High Voltage Reference Manual

HIGH VOLTAGE REFERENCE MANUAL
7/09 REV.2
T E C H N I C A L
R E S O U R C E S
TABLE OF CONTENTS
M A N U A L
SECTION 1
Frequently Asked Questions
ARC/SHORT CIRCUIT
INTERFACING
SAFETY
TECHNOLOGY/TERMINOLOGY
USAGE/APPLICATION
SECTION 2
Application Notes
AN-01
WHAT DO YOU MEAN; THE OUTPUT
IS GROUND REFERENCED?
AN-02
GROUND IS GROUND, RIGHT?
WHAT YOU NEED TO KNOW
AN-03
WHEN OVER SPECIFYING A POWER
1
1
2
3
6
8
8
SUPPLY CAN BE A BAD THING
IMPORTANT IN PROGRAMMING POWER SUPPLIES
TIMES EXPLAINED
AN-04
WHY SIGNAL TO NOISE RATIOS ARE
AN-05
HVPS OUTPUT FALL AND DISCHARGE
AN-06
JUST JUMPER THE EXTERNAL INTERLOCK?
WHY YOU REALLY SHOULDNT
AN-07
WHATS THE VOLTAGE RATING OF RG8-U
COAXIAL CABLE?
AN-08
HOW DO I CHANGE THE POLARITY
OF THE POWER SUPPLY?
SECTION 3
Articles
IEEE STD 510-1983 IEEE RECOMMENDED
PRACTICES FOR SAFETY IN HIGH VOLTAGE AND
HIGH POWER TESTING
SPECIFYING HIGH VOLTAGE POWER SUPPLIES

HIGH VOLTAGE POWER SUPPLIES FOR
ANALYTICAL INSTRUMENTATION
HIGH VOLTAGE POWER SUPPLIES FOR
ELECTROSTATIC APPLICATIONS
STANDARD TEST PROCEDURES FOR
HIGH VOLTAGE POWER SUPPLIES
COMPARATIVE TESTING OF SHIELD
TERMINATIONS OF HV CABLES
DESIGN AND TESTING OF A HIGH-POWER PULSED LOAD
ACCURATE MEASUREMENT OF ON-STATE
LOSSES OF POWER SEMICONDUCTORS
HIGHLY EFFICIENT SWITCH-MODE 100KV, 100KW
POWER SUPPLY FOR ESP APPLICATIONS
HIGH POWER, HIGH EFFICIENCY, LOW COST CAPACITOR
CHARGER CONCEPT AND DEMONSTRATION
SECTION 4
Glossary
10
11
11
12
14
18
22
26
37
42
47
52
58
64
T E C H N I C A L
R E S O U R C E S
FREQUENTLY ASKED QUESTIONS
ARC/SHORT CIRCUIT
page 1
The only limit to the magnitude of short circuit current is

the resistance in the series with the discharge circuit. All
Spellman supplies have built-in output limiting assemblies
that limit the instantaneous discharge current to a limited
level. The instantaneous short circuit current is determined
by the setting of the output voltage divided by the resistance that is in series with the discharge path. The amount
of time this discharge event is present(and its rate of
decay) is determined by the amount of capacitance and
resistance present in the discharged circuit.
Are your supplies current protected?
Virtually all of Spellman's supplies (with the exception of a

few modular proportional supplies) are "current protected."
Current protection is accomplished through the use of a
regulating current loop, otherwise known as current mode.
The current mode is programmed to a regulating level via
the front panel pot or the remote current programming signal. A current feedback signal is generated inside the supply that drives the current meter (if there is one) and the
remote current monitor signal. By comparing the current
feedback signal to the current program signal, the supply
can limit or regulated the output current to the desired
level. Even if a continuous short circuit is placed on the
output of the supply, the current mode will limit the output
current to the desired preset level.
When a short circuit is placed upon the output of a

supply, there is an instantaneous short circuit current.
Once the output capacitance has been discharged, additional output current can only come from the power generating circuitry of the power supply itself. To prevent this,
the power supply will sense the rise in output current due
to this short circuit condition and will automatically cross
over into current mode to regulate the output current to the
programmed present level.
Why is the short circuit repetition rate of my

load set-up important?
In summary, the instantaneous short circuit current is a

pulse of current that discharges the capacitance of the
supply, and the continuous short circuit current is the
current limit level set and controlled by the current mode
of the power supply.
How frequently a power supply is short circuited is an

important parameter to specify when selecting a supply
for a particular application.
As a rule of thumb, most of Spellman's supplies are designed to be short circuited at a 1 Hertz maximum repetition rate. This rating is dictated by the stored energy of the
output section of the supply, and the power handling capability of the internal resistive output limiter that limits the
peak discharge current during short circuiting. These
resistive limiters (that keep the instantaneous discharge
current to a limited level) thermally dissipate the stored
energy of the supply during short circuiting. If a supply is
arced at a repetition rate higher than it was designed for,
the resistive limiters in time, may become damaged due to
overheating. Brief bursts of intense arcing usually can be
handled, as long as the average short circuit rate is maintained at or below 1 Hertz.
INTERFACING
What kind of high voltage connector do you

use on your supplies?
While most Spellman supplies typically come with one of

two types of Spellman designed high voltage connector or
cable arrangements, many other industry standards
(Alden, Lemo,Kings, etc.) or custom cable/connectors
can be provided.
Many of our lower power modular supplies are provided

with a "fly wire" output cable. This output arrangement is a
length of appropriately rated high voltage wire that is permanently attached to the unit. This wire may be shielded
or non-shielded, depending on model. Catalog items come
with fixed lengths and non-standard lengths are available
via special order.
Supplies can be modified to enhance their short circuit

repetition rate by reducing their internal capacitance
and/or augmenting the power handling capability of the
resistive output limiting assembly. Please contact the
Sales Department for additional information.
What is the difference between instantaneous short

circuit current and continuous short circuit current?
The output section of a typical high voltage power supply

is capacitive, which causes it to store energy. When a
short circuit is placed on the output of a supply, the energy
stored in the capacitance of the multiplier is discharged.
SEC.1
Most higher power units, both modular and rack mounted,

are provided with a Spellman-designed and fabricated,
detachable, high voltage cable/connector assembly, often
referred to as a Delrin Connector. Typically a deep well
female connector is located on the supply and a modified
coaxial polyethylene cable/connector arrangement is provided. The coaxial cable's PVC jacket and braided shield
T E C H N I C A L
R E S O U R C E S
INTERFACING
page 2
controlled and programmed via a PC software interface

usually provided by the card vendor. Please contact our
Sales Department for additional information.
(continued)
is stripped back exposing the polyethylene insulation. The

length of the stripped back portion depends upon the voltage rating of the supply. A banana plug is attached to the
center conductor at the end of the cable and a modified
UHF or MS connector shell is used to terminate where the
stripped back portion of the cable ends. This allows for a
simple and reliable high voltage connection to be made to
the supply. Cables can be easily connected or detached
as required.
SAFETY
What is a safe level of high voltage?
Safety is absolutely paramount in every aspect of Spellman's high voltage endeavors. To provide the maximum
margin of safety to Spellman's employees and customers
alike, we take the stand that there is no "safe" level of high
voltage. Using this guideline, we treat every situation that
may have any possible high voltage potential associated
with it as a hazardous, life threatening condition.
Below is a photo of a typical detachable high voltage

Cable. Please contact the Sales Department for additional
information regarding special high voltage connector/cable
and custom lengths.
We strongly recommend the use of interlocked high voltage Faraday Cages or enclosures, the interlocking of all
high voltage access panels, the use of ground sticks to
discharge any source of high voltage, the use of external
interlock circuitry, and the prudent avoidance of any point
that could have the slightest chance of being energized to
a high voltage potential. The rigorous enforcement of
comprehensive and consistent safety practices is the
best method of ensuring user safety.
Typical Detachable High Voltage Cable
Can I program your supplies with a computer?
Where can I obtain information on high voltage

safety practices?
Yes, Spellman supplies can be programmed and

controlled with a computer.
Most of Spellmans newer product releases come

complete with our integrated SIC Option which provides
the ability to program the unit via RS-232, Ethernet or
USB protocols.
One of the most comprehensive publications regarding

high voltage safety practices is an excerpt from IEEE
Standard 510-1983 known as "The IEEE Recommended
Practices for Safety in High Voltage and High Power Testing." This information is available from Spellman in the
form of a printed document included in our "Standard Test
Procedures and Safety Practices for High Voltage Power
Supplies" handout. Please contact our Sales Department
for a copy.
Many of our standard products that do not show the SIC

Option as a possible offering on the data sheet, can in
some cases be modified to have the SIC Option added to
them. Please consult the Sales Department for details.
Supplies that can not be provided with the SIC Option can
still be computer controlled.
What is an "external interlock"? Why should I use it?
An external interlock is a safety circuit provided for customer use. Most interlock circuits consist of two terminals
provided on the customer interface connector. A connection must be made between these two points for the power
supply to be enabled into the HV ON mode. It is strongly
recommended that these interlock connections be made
via fail safe electro-mechanical components (switches,
contactors, relays) as opposed to semiconductor transistor
devices. If the power supply is already in the HV ON mode
and the connection is broken between these points, the
unit will revert to the HV OFF mode.
Virtually all of our products can be remote programmed

via an externally provided ground referenced signal. In
most cases 0 to 10 volts corresponds to 0 to full-scale
rated voltage and 0 to full-scale rated current. Output voltage and current monitor signals are provided in a similar
fashion. External inhibit signals and/or HV ON and HV
OFF functioning can be controlled via a ground referenced
TTL signal or opening and/or closing a set of dry contacts.
More detailed information regarding interfacing is provided
in the product manual.
There are several third-party vendors that sell PC interface cards that can act as an interface between the
signals detailed above and a PC. These cards can be
SEC.1
This simple circuit allows the customer to connect their

own safety interlock switch to the power supply. This
T E C H N I C A L
R E S O U R C E S
SAFETY
SEC.1
page 3
users, where a single specific usage needs to be

addressed in the most compact and cost effective
manner possible. These are guidelines, not rules.
(continued)
switch could be an interlock connection on a HV access

panel. In this way, if the panel was inadvertently opened,
the high voltage would be turned off, greatly reducing the
risk of bodily harm or physical injury. Spellman strongly recommends the use of interlock circuitry whenever possible.
Module
Rack
What is the difference between voltage

mode and current mode?
Voltage mode and current mode are the two regulating

conditions that control the output of the supply. Most applications call for a supply to be used as a voltage source.
A voltage source provides a constant output voltage as
current is drawn from 0 to full rated current of the supply.
In these applications, the power supply runs in voltage
mode, maintaining a constant output voltage while
providing the required current to the load. A voltage
source is generally modeled as providing a low output
impedance of the supply.
External Interlock
What is the difference between a modular

supply and a rack supply?
Modular supplies and rack supplies are the two generic

categories into which Spellman's standard products typically fall. These product categories were created and used
to help classify hardware. Additionally, Spellman provides
a variety of custom and OEM supplies that would not
adequately fit into either category.
Current mode works in a similar fashion, except it limits

and regulates the output current of the supply to the
desired level. When the supply runs in current mode, the
supply provides a constant current into a variety of load
voltage conditions including a short circuit. A current
source is generally modeled as providing a very high
output impedance of the supply.
Typically, rack mounted supplies are higher in power than

their modular counterparts; but this is a generalization, not
a rule. Rack mounted units usually operate off-line,
requiring AC input. Rack mounted units usually provide
full feature front panels, allowing quick and easy operator
use. Spellman's rack mounted supplies comply with the
EIA RS-310C rack-mounted standards.
These two regulating modes work together to provide

continuous control of the supply, but with only one mode
regulating at a time. These are fast acting electronic
regulating circuits, so automatic crossover between voltage mode to current mode is inherent in the design. With
the programming of the voltage mode and current mode
set points available to the customer, the maximum output
voltage and current of the supply can be controlled under
all operating conditions.
Modular supplies tend to be lower power units (tens to

hundreds of watts) housed in a simple sheet metal enclosure. Modular units that can operate off AC or DC inputs,
can be provided. OEM manufacturers frequently specify
modular supplies, knowing the elaborate local controls
and monitors are usually not included, thus providing a
cost savings. Customer provided signals, done via the
remote interface connector, usually accomplishes
operation, programming and control of these units.
When ease of use and flexibility is required, like in a

laboratory environment, rack mounted supplies are usually
preferred. Modular supplies tend to be specified by OEM
What is power control? When would it be used?
Power control, (a.k.a. power mode or power loop) is a

third control mode that can be added to a variety of
Spellman supplies to provide another means to control
and regulate the output of the supply. Voltage mode and
current mode are the primary controlling modes of most
units. Taking the voltage and current monitor signal and
inputting them into an analog multiplier circuit, creates a
T E C H N I C A L
R E S O U R C E S
SEC.1
page 4
(continued)
power feedback signal (voltage x current = power). Using

this feedback signal with an additional programmable
reference signal in conjunction with error amplifier
circuitry, a programmable power mode can be created.
Power control is typically used in two types of applications.

The less common application is where the power into a
load is the needed regulating parameter. A critical heating
requirement may have very specific regulated thermal
need. Using power mode, voltage and current limit levels
can be established, and power mode will provide constant
power to the load, immune from any impedance variations
from the load itself.
Floating Ground
What is solid encapsulation?
The more popular usage of a power mode is in the area

where a power source or load might be rated or capable of
more current at reduced voltage levels, but limited to a
particular power level. X-ray tubes frequently have this
type of capability. If the maximum voltage were multiplied
by this "increased current" capability, a power level above
the rated power level would result. Power mode can address this problem by limiting the power to the maximum
rated (or present) level.
Solid encapsulation, also referred to as "potting," is an insulation media used in a variety of Spellman's supplies.
The "output section" of a high voltage power supply can
operate at extremely high voltages. The design and packaging of the high voltage output section is critical to the
functionality and reliability of the product.
Solid encapsulation allows Spellman designers to miniaturize the packaging of supplies in ways that are unobtainable when utilizing air as the primary insulating media
alone. Improved power densities result, providing the
customer with a smaller, more compact supply.
What is floating ground?
The term floating ground (FG) is used to describe an option that allows for very accurate ground referenced load
current measurements to be made.
Additionally, solid encapsulation provides the feature of

sealing off a potted output section from environmental
factors. Dust, contamination, humidity and vibration typically will not degrade or affect the performance of an encapsulated high voltage output section. This is especially
important where a supply will operate in a harsh environment, or where a unit must operate maintenance free.
Whatever current flows out of the high voltage output of a

supply, must return via the ground referenced return path.
This current must return back to its original source, the
high voltage output section inside the supply.
The FG option isolates all of the analog grounds inside
the supply and brings them to one point: usually provided
on the rear of the power supply. If a current meter is
connected between this FG point and chassis ground,
the actual high voltage return current can be measured
in a safe ground referenced fashion.
Why is oil insulation used?
Spellman has invested in and developed the use of oil

insulation technology, giving its engineers and designers,
when appropriate, another method of high voltage packaging technology. Oil, as an insulating media has some
distinct advantages in particular situations. This capability
has been utilized in several of Spellman's MONOBLOCK
designs, where a power supply and an X-ray tube assembly have been integrated into a single unit. The results of
this integration include a reduction of the size and weight
of a unit, in addition to providing excellent heat transfer
characteristics and eliminating costly high voltage cables
and connectors.
Essentially, the analog grounds inside the supply are

"floated" up a few volts to allow for this measurement.
This option is only intended to allow for a ground referenced current measurement, so the actual maximum
voltage the internal analog ground "floats" to, is usually
limited to 10 volts maximum.
It is important to note that all control and monitoring circuitry are also floated on top of the FG terminal voltage.
Users of this option must provide isolation from the FG terminal to chassis ground. Higher voltages may be available
depending on the model selected. Please contact our
Sales Department for more information.
T E C H N I C A L
R E S O U R C E S
What is corona?
SEC.1
page 5
(continued)
Corona is a luminous, audible discharge that occurs when

there is an excessive localized electric field gradient upon
an object that causes the ionization and possible electrical
breakdown of the air adjacent to this point. Corona is characterized by a colored glow frequently visible in a darkened environment. The audible discharge, usually a subtle
hissing sound, increases in intensity with increasing output
voltage. Ozone, an odorous, unstable form of oxygen is
frequently generated during this process. Rubber is
destroyed by ozone, and nitric acid can be created if sufficient moisture is present. These items have detrimental
affects on materials, inclusive of electrical insulators.
A good high voltage design takes corona generation into
account and provides design countermeasures to limit the
possibility of problems developing. Spellman engineers
use sophisticated e-field modeling software and a Biddle
Partial Discharge Detector to ensure that each high voltage design does not have
excessive field gradients,
preventing partial discharge
and corona generation.
Resonant Inverter
What is a voltage multiplier?
A voltage multiplier circuit is an arrangement of capacitors

and rectifier diodes that is frequently used to generate
high DC voltages. This kind of circuit uses the principle of
charging capacitors in parallel, from the AC input and
adding the voltages across them in series to obtain DC
voltages higher than the source voltage. Individual voltage
multiplier circuits (frequently called stages) can be connected in series to obtain even higher output voltages.
Spellman has pioneered the use of voltage multiplier

circuits at extreme voltage and power levels. Spellman's
engineers have repeatedly broken limits normally associated with this type of circuit, as they continue to lead in the
development of this area of high voltage technology.
Corona and Breakdown
Corona
What is a resonant inverter?
A resonant inverter is the generic name for a type of high

frequency switching topology used in many of Spellman's
supplies. Resonant switching topologies are the next generation of power conversion circuits, when compared to
traditional pulse width modulation (PWM) topologies.
Resonant-based supplies are more efficient than their

PWM counterparts. This is due to the zero current and/or
zero voltage transistor switching that is inherent in a resonant supplies design. This feature also provides an additional benefit of eliminating undesireable electromagnetic
radiation normally associated with switching supplies.
High Voltage Multiplier
T E C H N I C A L
R E S O U R C E S
USAGE/APPLICATION
SEC.1
page 6
current can only flow out of the supply. Because the supply can't sink current, the charged output capacitance
needs to be bled off into the customer's load or some
other discharge path.
Positive polarity, negative polarity, reversible polarity; why is this important when I purchase a supply?
Our positive supplies actually do have a small amount of

"current sink" capability provided by the resistance of the
voltage feedback divider string, located inside the supply.
An extremely high value of resistance is necessary(typically tens or hundreds of meg-ohms, or even gig-ohms) so
the output capacitance will bleed off to zero volts, in seconds or tens of seconds in a "no load" condition. For this
reason, the fall time of our supplies are load dependent.
DC sources are polarity specific. Using earth ground as a

reference point, the output of a DC supply can be "X"
number of volts above ground (positive polarity) or "X"
number of volts below ground (negative polarity). Another
way of explaining this, is as a positive supply can source
(provide) current, while a negative supply can sink (accept) current. Applications that require DC high voltage
sources are polarity specific, so the polarity required must
be specified at the time of order.
How should I ground your supply?
Can I run your supplies at maximum voltage?

Maximum current? How much should I de-rate your
supplies?
Grounding is critical to proper power supply operation.

The ground connection establishes a known reference potential that becomes a baseline for all other measurements. It is important that grounds in a system are low
impedance, and are connected in such a way that if currents
flow through ground conductors they do not create voltage
level changes from one part of the system to another.
Spellman standard supplies can be run at maximum voltage, maximum current, and maximum power continuously
with no adverse affect on performance or reliability. Each
supply we sell is burned in at full rated voltage and full
rated current for a minimum of 12 hours. All of our supplies are designed to meet a set of Spellman Engineering
Design Guidelines that dictate all appropriate internal
component deratings. Designing to these guidelines provides a supply with more than adequate margins, so there
is no need to derate our supplies below our specifications.
The best way to minimize the possibility of creating voltage differences in your system grounding is to use ground
planes via chassis and frame connections. Since the
source of the high voltage current is the power supply, it is
recommended that it be the tie point for system grounds to
other external devices.
Can I get twice the current from your supply

if I run it at half voltage?
Most of our unmodified products (with the exception of

several X-ray generators) obtain maximum rated power at
maximum rated voltage and maximum rated current.
Where more current is needed at lower voltages, we can
provide a custom design for your particular application.
Please contact our Sales Department to see how we
can satisfy your requirement.
Why is the fall time of your supplies load dependent?
A high voltage power supply's output section is capacitive

by design. This output capacitance gets charged up to the
operating voltage. When the supply is placed in HV OFF
or standby (or turned off entirely) this charged output
capacitance needs to be discharged for the output voltage
to return back to zero.
Most high voltage output sections use diodes in their output rectification or multiplication circuitry. The diodes are
orientated to provide the required output polarity. A diode
only allows current to flow one way. In a positive supply,
Power Supply Grounding
The rear panel of the power supply should be connected

to this system ground in the most direct, stout manner
possible, using the heaviest gauge wire available, connected in a secure and durable manner. This ties the
chassis of the supply to a known reference potential.
It is important to understand most damage to HV power
supplies occur during load arcing events. Arcing produces
very high transient currents that can damage power sup-
T E C H N I C A L
R E S O U R C E S
USAGE/APPLICATION
(continued)
ply control circuitry (and other system circuitry) if grounding is not done properly. The product manual provides
more detailed information regarding grounding requirements. If you have any additional questions, please contact the Sales Department.
Can I float your supplies?
Spellman's standard products are for the most part, designed and intended for use as ground referenced power
supplies. That is, only one high voltage output connection
is provided, while the current return path is made via the
customer-provided ground referenced load return wiring.
This load return must be connected to a reliable earth ground
connection for proper operation and transient protection.
Many applications do exist, like ion beam implantation,
which require supplies to operate at reference voltages
other than earth ground. A supply of this nature is said to
"float" at some other reference potential. If your application requires a floating power supply, please contact our
Sales Department to review your requirement.
SEC.1
page 7
T E C H N I C A L
R E S O U R C E S
APPLICATION NOTES
page 8
A ground system starts with whatever you use as your

ground reference point. There are several that can be
used: cold water pipe, electrical service conduit pipe, electrical service ground wire, a building's steel girder framework, or the old fashioned ground rod. Whichever you use;
connect this point to the ground stud on the HVPS with a
short, heavy gauge wire and appropriate lug. Earth is the
universal reference point and by tying the HVPS to it in
this manner you will create a good reference point.
The next important ground connection that's needed is the
load return. Whatever current comes out of the HVPS (be
it continuous rated current or transient arc current) must
have a return path back to the power supply. This path
should be an actual physical wire; again of a short, heavy
type. With this connection the large transient arc currents
will travel in a known path, without influencing other
ground referenced equipment.
AN-01
What do you mean; the output is

"ground referenced"?
Most of Spellman's standard catalog products are termed

to be "ground referenced power supplies". A ground referenced power supply typically only has only one (1) rated
high voltage output connector. Internally the high voltage
multiplier return is referenced to the grounded chassis of
the unit. This chassis is referenced to "house ground" in
the customer's system via the safety ground wire in the
power cable and a separate customer provided system
ground connection. With the output of the supply ground
referenced it is easy to sample the output voltage and current to obtain the feedback signals needed to regulate the
supply. A high impedance, ground referenced, high voltage feedback divider monitors the output voltage, while a
ground referenced current feedback resistor placed in series with the multiplier return monitors the output current.
Just a point of clarification: the "3rd green ground wire" in

the AC power line cord is NOT an adequate system
ground. This wire is a safety ground not intended to be
used as part of a grounding system. A washing machine
typically has a metal chassis. If an AC power wire popped
off inside and touched against the chassis you wouldn't
want to open the lid and get shocked. Here, the "3rd wire"
grounds the chassis, preventing a shock by bypassing the
current to earth. That is its function; to be only a redundant
safety ground. Don't rely on this connection as part of your
system ground scheme.
With the customer's load being referenced to ground the

circuit is complete. All measurements made with regards
to the power supply utilize earth ground as the reference
potential. Ground referencing a power supply simplifies its
design, and fabrication. All programming and monitoring
signals are also ground referenced, simplifying operation
Ground referenced power supplies can not in their native
form be "stacked one on top of another" to obtain higher
output voltages. All output circuitry is referenced to
ground, preventing it from being connected to any other
voltage source or reference potential.
Connect all additional system ground references to the

main grounding point of the high voltage power supply. Be
it a "star" ground system or a ground frame/plane system,
attached the ground connection to the power supply main
grounding point. Following these recommendations will
help create a proper functioning grounding system.
AN-02
Ground is ground, right? Well, not always.

What you need to know.
AN-03
You wouldnt use a pickaxe for dental surgery: When

over specifying a power supply can be a bad thing.
Ground is one of those "ideal" things like the "ideal switch"

that's spoken about in engineering school. An ideal switch
has all the good characteristics (no losses, zero switch
time, etc) and no bad ones. The truth is, ground is only as
good as you make it, and only keeps its integrity if you do
the right thing.
It's much easier to start from scratch and create a good

ground system than to try to fix a bad one. Grounding
problems can be difficult to isolate, analyze and solve.
Here are a few tips on creating a good ground system that
will benefit both your high voltage power supply and the
rest of your system.
SEC.2
Selecting the right power supply for the task at hand will
reward you in several ways like: reduced size, weight, cost
and superior performance. Over specifying and purchasing "more supply than you need" can actually result in
degraded system performance in some circumstances.
All Spellman power supplies are designed, built and tested

at their full rated output voltage and current. We have
applied the appropriate component deratings for reliable
long term operation at full rated voltage and current. No
additional deratings of our power supplies are required.
T E C H N I C A L
R E S O U R C E S
APPLICATION NOTES
AN-03
SEC.2
page 9
Lets look at two example units, where 0 to 10 volts of voltage programming equates to 0 to 100% of output voltage.
The first unit is an SL100P300 (100kV maximum) and the
second unit is an SL1P300 (1kV maximum).
(continued)
If you need 30kV, buy a 30kV unit and run it at 30kV; it's
what it was designed to do. The same goes for current
and power. You will get the most bang for the buck buying
a supply that closely fits your requirements. If you can afford a larger, heavier and more expensive supply there is
nothing wrong with having a bit more capacity, but, over
specifying is NOT required to get reliable operation. Minor
over specifying can result in additional weight, size and
cost. Gross over specifying can actually degrade system
performance.
If a rather low output voltage of 100 volts was desired, lets

look at the level of programming voltage each unit requires.
SL100P300
(100/100,000) (10) = 10mV
SL1P300
(100/1000) (10) = 1 volt
The SL100P300 needs a programming signal of 10mV,

while the SL1P300 needs a programming signal of 1 volt
to achieve the same 100 volt output.
You wouldn't use a 4 inch wide exterior house paint brush

to touch up delicate interior wooden trim molding. A large
brush is great for quickly applying a lot of paint to a big
area, but a smaller brush allows better application and
control when painting smaller items. Size the tool for the
intended job to get the best results.
Noise is present in most electrical systems; its the low

level background signal that is due to switching regulators,
clock circuits and the like. Ideally zero noise would be desired, but some amount is present and must be dealt with.
In a power supply like the SL Series 25mV of background
noise on the analog control lines is not uncommon. Ideally
we would like to have the programming signal as large as
possible, so the noise signal has the least amount of influence. Lets see how that noise affects the signals of our
two example power supplies.
Power supplies are similar. A 30kV supply can operate

down at 250 volts, but when running at less that 1% of its
rated output, it can be somewhat hard to control with great
resolution. A 500 volt or even 1kV rated maximum output
supply would more adequately address this requirement.
SL100P300
Signal = 10mV
Noise = 25mV
s/n ratio: signal is
smaller than noise
None of our supplies have any "minimum load requirements". But keep in mind if excellent low voltage or low
power operation is required select a supply with maximum
ratings that are close to your needs. It's easier to obtain
precision operation when the power supply is properly
scaled and selected for its intended usage. If not, issues
like miniscule program and feedback signals, signal to
noise ratios, feedback divider currents can make operating
a supply at very small percentages of it's maximum rated
output very difficult.
SL1P300
Signal = 1000mV (1 volt)
Noise = 25mV
s/n ratio: signal is 40X
larger than noise
Its easy to see that getting a stable, repeatable 100 volt

output from the SL100P300 will be quite difficult, while this
is easy to do with the SL1P300.
When low output voltages are needed think about the
programming signals required and how they compare to
the system noise levels. Doing so will provide a stable,
repeatable output where noise has minimal effect.
AN-04
How low can you go?

Why signal to noise ratios are important in
programming high voltage power supplies.
AN-05
No, you touch it. HVPS output fall and discharge

times explained.
Virtually all Spellman power supplies are programmable;

usually a 0 to 10 volt ground referenced analog programming signal is proportional to 0 to 100% of full scale rate
voltage and/or current. Modular supplies typically only accept a remotely provided signal, while rack units also have
front panel mounted multi-turn potentiometers to provide
local programming capability.
When working with high voltage power supplies knowing

about output fall and discharge times can be helpful. Consider this information as only providing additional details
on power supply functionality. This application note by itself is not adequate "safety training" for the proper setup
and use of a HVPS. Please refer to the complete safety
information provided with our products.
T E C H N I C A L
R E S O U R C E S
APPLICATION NOTES
AN-05
SEC.2
page 10
AN-06
(continued)
"Just jumper the external interlock"?

Why you really shouldn't.
Typically, high voltage is created by controlling an inverter

that feeds a step up transformer which is connected to a
voltage multiplier circuit. This multiplier circuit (an arrangement of capacitors and diodes) uses the principle of
charging and discharging capacitors on alternate half
cycles of the AC voltage, where the output is the sum
of these capacitor voltages in series. By definition, the
voltage multiplier circuit is capacitive in nature and has
the ability to store and hold charge.
Many Spellman high voltage power supplies come with an

external interlock feature. Typically the external interlock is
provided by means of two signal connections on the rear
panel terminal block or interface connector. This feature
provides the user the ability to shut off and prevent the
generation of high voltage in a fail safe manner. This external interlock circuitry can easily be incorporated into the
user's setup to provide an additional level of operator safety.
For the sake of efficiency, any internal current paths to

ground are minimized. Typically the only resistive path
connecting the output of the supply to ground is the high
impedance voltage feedback divider string. This feedback
divider generates the low level, ground referenced, voltage
feedback signal used to control and regulate the supply.
In most cases the current of the relay coil that is used to

latch the power supply into the HV ON mode is routed out
to, and back from, the rear panel external interlock points.
This is usually a low voltage relay coil; 12Vdc or 24Vdc
with current in the range of tens of milliamps. The two
external interlock points must be connected together with
a low impedance connection to allow the power supply
to be placed into, (and to continue to operate in) the HV
ON mode.
Due to the orientation of the diodes in the multiplier assembly, a positive polarity supply can only source current;
it has no ability to sink current. So the feedback divider
string becomes the only discharge path for the output during a "no-load" condition. Let's look at a typical unit's value
of multiplier capacitance and feedback divider resistance
to see what kind of no load RC discharge time constants
we're talking about.
Opening this connection will prevent the supply from being

placed in the HV ON mode. Additionally, if the unit was actively running in the HV ON mode, open this connection
would cause the power supply to revert to the HV OFF
mode. The external interlock is the best method of controlling the power supply output with regards to safety, other
than disconnecting the power supply from its input power
source.
SL60P3000
60kV, 0- 5mA, 300 watts
C multiplier = 2285pF R feedback = 1400M
RC = (2285pF) (1400M) = 3.199 seconds
5 RC time constants required to approach zero (1.2%)
(5) (3.199 seconds) = 15.995 seconds
Typically our power supplies are shipped with the two

external interlock connections jumpered together to allow
quick and easy operation of the supply. Leaving the unit
configured in this manner does indeed work, but it
bypasses the external interlock function.
The above example illustrates how under a no load condition it can take considerable time for the output to discharge. If an external load is left connected to the supply's
output, the discharge time constant can be shortened considerably. For this reason HVPS fall times are termed to
be "load dependent". Keep this in mind when working with
your next HVPS.
Spellman recommends that any exposed high voltage

potential be isolated from contact through the use of
appropriate physical barriers. High voltage cages or
enclosures should be used to protect operators from
inadvertent contact with potentially lethal voltages.
Doors and/or access panels of these cages or enclosures
should have a normally open interlock switch installed on
them such that the switch is in the closed state only when
the door or panel is in the secured position. Opening the
door or panel will revert the power supply to the HV OFF
mode, and prevent the supply from being placed in the HV
ON mode until the door or panel is properly secured.
10
T E C H N I C A L
R E S O U R C E S
APPLICATION NOTES
page 11
AN-08
AN-07
What's the voltage rating of RG8-U coaxial cable?
How do I change the polarity of the power supply?
Output cable and connectors are not trivial items for power
supplies where output voltages can be 100,000 volts or
higher. The cables and connectors used must function
together as a system to safely and reliably access and
provide the power supplies output for customer usage.
How do I change the polarity of the power supply?

Most high voltage power supplies use a circuit called a
voltage multiplier to create the desired high voltage output.
This basic multiplier circuit is shown below in the simplified
power supply block diagram:
In many high voltage power supply applications, a

shielded polyethylene coaxial cable is used. Polyethylene
cables provide excellent high voltage dielectric isolation
characteristics in a small but robust form factor. The shield
conductor provided in a coaxial cable functions as a
"Faraday Shield" for the center conductor of the cable that
is referenced to the high voltage potential. If any breakdown in the main insulator occurs, the high voltage current
will be bypassed to the grounded shield conductor that
surrounds the main insulator. This inherent safety feature
is one benefit of using a coaxial high voltage output cable.
IMAGE HIGH VOLTAGE POWER SUPPLY
RG8-U has long been used as a high voltage output cable

in the high voltage industry. There is a variation of RG8-U
that utilizes a solid polyethylene core. Specifications for
this cable do not specify actual "high voltage" ratings,
since this cable was not designed and fabricated with high
voltage usage in mind. So the reality is, there are no high
voltage ratings for RG8-U. Over the years others in the HV
industry have used this cable at 20kV, 30kV and even
higher voltages. Spellman does use RG8-U cable, but
limits it usage to applications where the maximum voltage
that will be applied to the cable is 8kV or less.
Simplified Schematic Diagram of

a High Voltage Power Supply
The multiplier circuit is comprised of an arrangement of

capacitors and diodes. The orientation of the diodes will
determine the output polarity of the unit. In the example
above, the diodes shown would create a positive output
polarity with respect to ground. If each diode was reversed
in orientation, the multiplier would generate a negative
output voltage with respect to ground.
The example above only shows a two stage, half-wave
multiplier; using a total of four diodes. Full-wave multiplier
stages are more efficient and use additional capacitors
and twice as many diodes. To generate the high voltages
typical of a Spellman supply, many multiplier stages are
connected in series. If a twelve stage, full wave multiplier
was made, a total of 48 diodes would be required.
For voltages above 8kV where a coaxial polyethylene

cable is desired, Spellman uses cables specifically
designed and manufactured for high voltage usage.
These cables are of the same general design; as described above but the insulating core material diameter
has been increased appropriately to obtain the desired
dielectric insulating capability required. Frequently higher
voltage versions of these cables utilize a thin semiconductor "corona shield". This corona shield is located between
the metallic center conductor and the main polyethylene
insulating core. This corona shield helps equalize the
geometric voltage gradients of the conductor and by doing
so reduces the generation of corona.
A high voltage cable and connector system can only be as

good as the materials used to make it. Using cables that
are designed, specified and tested specifically for high
voltage usage assures that these materials are used
within their design guidelines.
SEC.2
Typically the capacitors and diodes used to fabricate a multiplier assembly are soldered directly to a single or sometimes
several printed circuit boards. Frequently this assembly is
encapsulated for high voltage isolation purposes.
11
To simplify the process of reversing the polarity (like in the

instance of the SL Series) a second "opposite polarity"
multiplier is provided above 8kV when reversibility is
required. Exchanging the multiplier is a simple task needing only a screwdriver and few minutes of time. Modular
style units due to their simplified design, are typically not
capable of having their polarity changed in the field.
T E C H N I C A L
R E S O U R C E S
ARTICLES
IEEE Std 510-1983 IEEE Recommended

Practices for Safety in High Voltage and
High Power Testing
SEC.3
page 12
iEEE Std 510-1983 IEEE Recommended Practices for Safety
TEST AREA SAFETY PRACTICES
Appropriate warning signs, for example,

DANGER HIGH VOLTAGE, should be posted on
or near the entrance gates.
by The Institute of Electrical and Electronics Engineers
Insofar as practical, automatic grounding devices

should be provided to apply a visible ground on the
high-voltage circuits after they are de-energized. In
some high-voltage circuits, particularly those in which
elements are hanged from one setup to the next, this
may not be feasible. In these cases, the operator
should attach a ground to the high-voltage terminal
using a suitably insulated handle. In the case of several
capacitors connected in series, it is not always sufficient
to ground only the high-voltage terminal. The exposed
intermediate terminals should also be grounded. This
applies in particular to impulse generators where the
capacitors should be short-circuited and grounded be
fore and while working on the generator.
SCOPE
Excerpts from IEEE Standard 510-1983 have been listed

in this section in order to caution all personnel dealing with
high voltage applications and measurements and to provide recommended safety practices with regard to electrical hazards.
Considerations of safety in electrical testing apply not only
to personnel but to the test equipment and apparatus or
system under test. These recommended practices deal
generally with safety in connection with testing in laboratories, in the field, and of systems incorporating high voltage
power supplies, etc. For the purposes of these recommended practices, a voltage of approximately 1,000 volts
has been assumed as a practical minimum for these types
of tests. Individual judgement is necessary to decide if the
requirements of these recommended practices are applicable in cases where lower voltages or special risks are
involved.
Safe grounding of instrumentation should take precedence over proper signal grounding unless other precautions have been taken to ensure personnel safety.
CONTROL & MEASUREMENT CIRCUITS
Leads should not be run from a test area unless they are
contained in a grounded metallic sheath and terminated in
a grounded metallic enclosure, or unless other precautions have been taken to ensure personnel safety. Control
wiring, meter connections, and cables running to oscilloscopes fall into this category. Meters and other instruments with accessible terminals should normally be
placed in a metal compartment with a viewing window.
All ungrounded terminals of the test equipment or apparatus under test should be considered as energized.
Common ground connections should be solidly connected to both the test set and the test specimen. As a
minimum, the current capacity of the ground leads
should exceed that necessary to carry the maximum
possible ground current. The effect of ground potential
rise due to the resistance and reactance of the earth
connection should be considered.
Temporary Circuits
Temporary measuring circuits should be located completely within the test area and viewed through the
fence. Alternatively, the meters may be located outside
the fence, provided the meters and leads, external to
the area, are enclosed in grounded metallic enclosures.
Precautions should be taken to prevent accidental

contact of live terminals by personnel, either by shield
ing the live terminals or by providing barriers around
the area. The circuit should include instrumentation for
indicating the test voltages.
Temporary control circuits should be treated the same

as measuring circuits and housed in a grounded box
with all controls accessible to the operator at ground
potential.
Appropriate switching and, where appropriate, an

observer should be provided for the immediate de-energization of test circuits for safety purposes. In the case
of dc tests, provisions for discharging and grounding
charged terminals and supporting insulation should also
be included.
SAFETY RULES
High Voltage and high-power tests should be performed

and supervised by qualified personnel.
12
A set of safety rules should be established and enforced for the laboratory or testing facilities. A copy
of these should be given to, and discussed with, each
person assigned to work in a test area. A procedure for
periodic review of these rules with the operators should
be established and carried out.
T E C H N I C A L
R E S O U R C E S
ARTICLES
SEC.3
page 13
iEEE Std 510-1983 IEEE Recommended Practices for Safety
SAFETY INSPECTION
HIGH-POWER TESTING
A procedure for periodic inspection of the test areas

should be established and carried out. The recommendations from these inspections should be followed by
corrective actions for unsafe equipment or for practices
that are not in keeping with the required regulations.
High-power testing involves a special type of high-voltage

measurement in that the level of current is very high. Careful consideration should be given to safety precautions for
high-power testing due to this fact. The explosive nature of
the test specimen also brings about special concern relating to safety in the laboratory.
NOTE: A safety committee composed of several operators

appointed on a rotating basis has proven to be effective,
not only from the inspection standpoint but also in making
all personnel aware of safety.
Protective eye and face equipment should be worn by all

personnel conducting or observing a high-power test
where there is a reasonable probability that eye or face
injury can be prevented by such equipment.
GROUNDING & SHORTING
NOTE: Typical eye and face hazards present in high-power

test areas included intense light (including ultraviolet), sparks,
and molten metal.
The routing and connections of temporary wiring should

be such that they are secure against accidental interruptions
that may create hazard to personnel or equipments.
Safety glasses containing absorptive lenses should be

worn by all personnel observing a high-power test even
when electric arcing is not expected. Lenses should be
impact-resistant and have shade numbers consistent with
the ambient illumination level of the work area but yet capable of providing protection against hazardous radiation
due to any inadvertent electric arcing.
Devices which rely on a solid or solid/liquid dielectric for insulation should preferably be grounded and short-circuited
when not in use
Good safety practice requires that capacitive objects be
short-circuited in the following situations:
Any capacitive object which is not in use but may be in the

influence of a dc electric field should have its exposed
high-voltage terminal grounded. Failure to observe this
precaution may result in a voltage included in the capacitive object by the field.
GENERAL
All high-voltage generating equipment should have a

single obvious control to switch the equipment off under
emergency conditions.
Capacitive objects having a solid dielectric should be

short-circuited after dc proof testing. Failure to observe
this precaution may result in a buildup of voltage on the
object due to dielectric absorption has dissipated or until
the object has been reconnected to a circuit.
All high-voltage generating equipment should have an indicator which signals that the high-voltage output is enabled.
All high-voltage generating equipment should have provisions for external connections (interlock) which, when
open, cause the high-voltage source to be switched off.
These connections may be used for external safety interlocks in barriers or for a foot or hand operated safety
switch.
NOTE: It is good practice for all capacitive devices to

remain short-circuited when not in use.
Any open circuited capacitive device should be short-circuited and grounded before being contacted by personnel.
The design of any piece of high-voltage test equipment

should include a failure analysis to determine if the failure
of any part of the circuit or the specimen to which it is connected will create a hazardous situation for the operator.
The major failure shall be construed to include the probability of failure of items that would be overstressed as the
result of the major failure. The analysis may be limited to
the effect of one major failure at a time, provided that the
major failure is obvious to the operator.
SPACING
All objects at ground potential must be placed away from

all exposed high voltage points at a minimum distance of 1
inch (25.4 mm) for every 7,500 Volts, e.g. 50 kV requires a
spacing of at least 6.7 inches (171 mm).
Allow a creepage distance of 1 inch (25.4 mm) for every

7,500 Volts for insulators placed in contact with high
voltage points.
13
T E C H N I C A L
R E S O U R C E S
ARTICLES
Specifying High Voltage Power Supplies
SEC.3
page 14
Compared with line frequency operation, high frequencies

offer the following advantages in regulated high voltage
power supplies:
by Derek Chambers and Cliff Scapellati
INTRODUCTION
Smaller size and weight

Faster response time
Lower stored energy
Higher efficiency
In specifying a regulated high voltage power supply for a

particular application, it is important to bear in mind that
recent advances in power supply technology have made
the latest designs smaller, lighter, more efficient than was
possible just a few years ago. New designs generally operate at high frequencies in the range of 20kHz to 100kHz,
and industry-wide, have virtually replaced all units operating at line frequency, even at high power levels.
All high voltage power supplies must be operated by
personnel familiar with the dangers of high voltage. High
voltage sources can be lethal! A general guideline for
Safety Practices is found in IEEE Standard 510-1983
"Recommended Practices for Safety in high voltage and
high power testing."
High-voltage supplies such as this multiple-output model use more

efficient and higher-performance components and power conversion
techniques to reduce weight and improve performance.
TECHNOLOGY
The two primary factors which have led to these

developments are:
The heart of any high frequency power supply is the oscillator (or inverter) used to drive the output transformer. The
specific designs used in the high voltage power supply industry are too numerous to cover in this article since each
manufacturer has developed his own proprietary power
switching circuits. However, there is one factor, unique to
high voltage power supplies, that must be considered in
the choice of the oscillator or inverter topology. Specifically, the capacitance which exists across the secondary
winding of the step-up transformer must be isolated from
being reflected directly across the power switching semiconductors. This isolation can be achieved in a number of
ways, including:
The availability of key power components which

have low losses while operating at high frequency
The development of advanced resonant power
conversion techniques
Key Power Components include:

Faster switching devices (e.g. transistors, power
MOSFETS, IGBTs, SCRs)
Low loss ferrite and powdered iron core materials for

choke and transformer cores
Capacitors with low dissipation factors
Using a flyback circuit
Ultra fast rectifiers which have a low

forward voltage drop
Using an inductor or a series resonant circuit between the switching devices and the transformer
Advanced Conversion Techniques include:

Zero current switching series and parallel resonant
inverters (discontinuous mode);
Including sufficient leakage inductance between the

primary and secondary windings of the transformer
Operating as a self resonant oscillator
Zero voltage switching LCC resonant inverters

(continuous mode)
Soft switching and phase controlled

resonant inverters
Quasi-resonant flyback and push-pull inverters
14
The choice of oscillator topology is also influenced by the

power level of the supply. For instance, a low power unit
for a photomultiplier application could use a flyback or self
resonant oscillator, while higher power models (e.g. over a
kilowatt) would be more likely to use a driven inverter
feeding the output transformer through an inductor or a
series resonant circuit. The transformer may also be designed to form part of the resonant inverter power circuit.
T E C H N I C A L
R E S O U R C E S
ARTICLES
SEC.3
page 15
Properly designed resonant converter designs offer the

following desirable characteristics:
Zero current switching, which improves efficiency

and minimize the switching losses in the high power
switching devices
Sinusoidal current waveforms in the power inverter
circuit, which greatly reduce RFl interference
normally associated with pulse width modulation
techniques
Higher-power high-voltage supplies, like Spellman's series SL which are

rated up to 1,200W, operate from ac line power.
Simple paralleling of the supplies to obtain higher

output power
Inherent current limiting and short circuit protection

of series resonant inverters
INPUT VOLTAGE
The input power source specified for a particular model is

determined by a number of factors including the output
power capability of the supply and the form of power available in the application. In general, low power high voltage
supplies having outputs between 1W and 60W use a dc
input voltage of 24V or 28V, while higher power units operate from the ac power line.
SPECIFICATION CONSIDERATIONS
Probably the most common mistake engineers make in

defining a high voltage power supply is to over specify the
requirements for output power, ripple, temperature stability, and size. Such over specification can lead to unnecessarily high cost, and can also lower reliability due to
increased complexity and greater power density. If a particular parameter in the catalog specification is inadequate
for the application, the factory should be consulted.
DC Input
In many OEM applications, the high voltage supply is
just one part of an electronic system in which dc
power sources are already available (e.g. 24Vdc,
390Vdc). These existing dc supplies can also be
used as the input power source for a high voltage
supply. This arrangement is convenient and economical for modular high voltage supplies operating
at low power levels.
UNDERSTANDING SPECIFICATION PARAMETERS
The specifications provided by the power supply manufacturer generally include information on the input and output
voltages, the output regulation, ripple, and output stability.
Often, more detailed information would be useful to the
user. In the following sections, power supply parameters
are discussed in greater detail than is normally possible
on a standard data sheet, and includes definitions and
descriptions of requirements encountered by users of
high voltage power supplies.
AC Input
Most high power modules over 100W, and rack
mounted models are designed for operation from an
ac line source. These power supplies are designed to
accept the characteristics of the power line normally
available at the location of the user, and these can
vary significantly in different parts of the world.
In the United States and Canada, the standard single
phase voltage is 115/230Vac at 60Hz, while in Continental
Europe and in many other parts of the world, the standard
voltage is 220Vac at 50Hz. In the UK, the standard is
240Vac at 50 Hz , while in Japan the voltage is normally
100V at 50 or 60Hz. Most power supplies include transformer taps to cover this range, while some new designs
cover the range 90Vac to 130Vac and 180Vac to 260Vac
without taps. All countries in the European Economic
Community will eventually standardize at 230V at 50Hz.
The specification parameters are covered in

the following order:
Input Voltage
Output Voltage
Output Current
Ripple
Stability
Stored Energy
Pulsed Operation
Line Regulation
Load Regulation
Dynamic Regulation
Efficiency
15
Power Factor correction and universal input at power

levels below 3kW can be specified for most off-the-shelf
high voltage power supplies. Higher power units require
custom engineering.
T E C H N I C A L
R E S O U R C E S
ARTICLES
OUTPUT VOLTAGE
page 16
The high frequency ripple may generally be reduced by

adding capacitance across the output. On the other hand,
when there is a fast response time requirement, the value
of output capacitance may have to be reduced. In critical
cases, the trade off between slew rate and ripple should
be worked out between the customer and the manufacturer of the power supply.
High voltage power supplies are generally designed for

continuous operation at the maximum output voltage
specified in the data sheet. Laboratory bench models and
high power rack units are normally adjustable over the
complete voltage range from zero to the maximum specified output voltage. In these models, output voltage is indicated on either digital or analog meters, as specified.
Modular supplies, on the other hand, may have either a
preset output voltage, or a narrow adjustment range, and
include monitor terminals instead of meters for measuring
the voltage. It is not generally cost effective to specify a
power supply with an output voltage greater than 20%
over the maximum voltage actually needed in a particular
application.
Line frequency ripple:

When operating from an ac input source, line frequency ripple can represent a significant part of the
total peak to peak ripple. Typically, the power supply
is designed to have equal amounts of high frequency
and line frequency ripple when operating at full output
power. It should be noted that, in most designs, the
magnitude of the line frequency ripple is attenuated
and controlled by feedback in the regulation circuits,
which normally have bandwidths to include the line
ripple frequency.
OUTPUT CURRENT
Switching frequency ripple:

In regulated supplies operating from a dc input, line
frequency ripple does not exist, and the ripple frequency is simply related to the switching or oscillator
frequency of the supply. To reduce switching frequency output ripple, additional filtering components,
or sometimes electronic ripple canceling circuits, may
be used. When filtering components, such as shunt
capacitors or series resistors or inductors, are added
to reduce the ripple, they introduce a delay in the
control loop circuits which adversely affects the response time of the supply to changes in input or output conditions. The values of the components which
control the phase of the signal in the feedback loop
are then changed at the factory to maintain stable
operation.
Power supplies are normally designed for continuous

operation at the full current specified in the data sheet.
Current limiting is normally built into the design to prevent
overload current from increasing beyond about 110% of
the rated maximum value of output current. Overload trip
out can usually be specified to disable the power supply
when the normal output current is exceeded. Current regulation is available on most high power racks and modules. This allows the output current to be controlled by a
front panel potentiometer or from a remote source, and
provides automatic crossover to voltage regulation when
the load current is lower than the programmed value.
RIPPLE
If an application requires particularly small values of either

high frequency or line frequency ripple, it is usually possible to provide a lower ripple at one of these frequencies at
the expense of increasing the ripple at the other. In these
special cases, the requirements should be discussed with
the factory before an order is placed.
Ripple may be defined as those portions of the output voltage that are harmonically related to both the input line
voltage and the internally generated oscillator frequency.
In high frequency switching designs it is the combined result of two frequencies, namely, the line frequency- related
components and the switching frequency related components. Total ripple is specified either as the rms, or the
peak-to-peak value of the combined line frequency and
oscillator frequency components, and is normally expressed as a percentage of the maximum output voltage.
The amount of ripple that can be tolerated in different

applications varies from extremely low values (e.g. less
than 0.001% peak to peak in photomultiplier, nuclear instrumentation and TWT applications) to several percent
when the output can be integrated over time, such as in
precipitators and E-beam welding.
SEC.3
STABILITY
The following factors affect the output stability of a

regulated high voltage power supply:
Drift in the reference voltage;

Offset voltage changes in the control amplifiers;
Drift in the voltage ratio of the feedback divider;
Drift in the value of the current sense resistor.
16
All these variations are a function of temperature. Stability

in a properly chosen reference device is generally less
than 5ppm, and offset errors can be virtually eliminated by
careful choice of the control amplifier. This leaves the volt-
T E C H N I C A L
R E S O U R C E S
ARTICLES
age divider and the current sense resistor as the critical

items affecting stability in the output voltage and current.
page 17
The first category includes pulsed radar applications

in which narrow pulses, having durations in the microsecond range, are generated at typical repetition
rates between 500Hz and 5kHz.
Since these components are sensitive to temperature variations, they are selected to operate at a fraction of their
power capability, and are located away from hot components. However, as the power supply warms up and the
ambient temperature around the components increases,
there are small changes in the ratio of the voltage divider
and the value of the current sense resistor which could affect stability.
The values for stability are usually given after a specified
warm-up period (typically 1/2 hour). Good stability is
achievable by using a divider with a low value of temperature coefficient, although this becomes more costly.
Compact high power module delivers to 350 watts CW or 600 watts

pulse for projection television and CRT testing. 1kV to 70kV with voltage
and current programming and monitoring.
STORED ENERGY
The stored energy at the output of a high voltage power

supply can be dangerous to operating personnel, particularly at the higher voltages since its value is a function of
the square of the voltage and the value of the capacitance
across the output. Certain types of loads, such as X-ray
tubes, are also easily damaged by excessive stored energy in the high voltage power supply when an arc occurs.
With power supplies operating at high frequency rather
than at line frequency, much smaller values of smoothing
capacitance can be used, and the dangers of electrocution
are thereby reduced. However, it should be noted that low
ripple power supplies which include additional filtering capacitance across the output have correspondingly higher
amounts of stored energy. Compared with a power supply
operating at line frequency, a switching supply operating
at 60kHz could have a fraction of the stored energy of an
equivalent line frequency supply, since the value of the
output capacitance could be reduced by 1000.
The second category covers a broader range of applications such as pulsed electromagnet supplies or cable testing where most of the pulse load current is still provided
by a capacitor connected across the output. Some modifications to the output and control circuits are usually
needed for reliable operation in these applications, and
the details of the load characteristics should be discussed
with the factory to ensure reliable operation in the customer's system.
The third category requires a power supply specifically
designed to provide more current than its average rated
value for relatively long periods. Typical applications are
medical X-ray systems, lasers and high voltage CRT
displays. It is essential that the actual load conditions are
completely specified by the user before placing an order.
PULSED OPERATION
LINE REGULATION
While some power supplies are designed for dc operation,

others can be used in pulsed power applications. In most
cases, an energy storage capacitor located inside or external to the supply provides the peak pulse current, and
the power supply replaces the charge between pulses.
The supply operates in the current mode during the pulse
and recharging parts of the cycle, and returns to the voltage mode before the next load current pulse.
Pulsed loads generally fall into one of three categories:
Line regulation is expressed as a percentage change in

output voltage for a specified change in line voltage,
usually over a 10% line voltage swing. Measurement is
made at maximum output voltage and full load current
unless otherwise stated. Line regulation of most high
voltage power supplies is better than 0.005%.
LOAD REGULATION
Very narrow pulses (1usec to 10usec), with a duty

ratio of 0.01% to 1%
Longer pulses (100usec to 1msec),
with a duty ratio between 0.05% and 0.2%
Very long pulses (50msec to 5sec),

with a duty ratio between 0.1% and 0.5%
SEC.3
17
Load Regulation is specified at full output voltage and

nominal line voltage and is expressed as a percentage
change in output voltage for a particular load current
change, usually no load to full load. Typical load regulation
of most high voltage supplies is better than 0.01%.
T E C H N I C A L
R E S O U R C E S
ARTICLES
can provide rectification and filtering for AC signals,

filtering for DC signals and circuit protection. Also,
auxiliary power sources to power the high voltage
power supply control circuits are typically part of the
power input stage responsibilities. It is critical for the
instrument designer to understand the input circuit
configurations. The power input requirements will
affect overall instrument design, customer requirements, and even regulatory requirements.
by Cliff Scapellati
Power supply requirements for Analytical Instrumentation

are as varied as the applications themselves. Power
supply voltages ranging from 3 volts to 300,000 volts can
be found within a given instrument. While most voltage
requirements can be satisfied with "off the shelf" products,
the high voltage requirements are usually addressed by a
custom design for a specific application. Custom designed
high voltage power supplies can be found in instruments
for spectroscopy, capillary electrophoresis, mass spectrometry, electrospray, lasers, spectrometers, X-ray
diffraction, X-ray fluorescence, and many other analytical
imaging and process applications.
Each application of High Voltage Power will require careful

attention to critical variables. Voltage ripple, long and short
term stability, repeatability and accuracy are important f
actors in the consideration of reliable scientific data. Also,
as analytical instrumentation finds its way into production
process control, reliability and quality are equally important in the considerations for high voltage power supply
specification.
Fig. 1
Basic High Voltage Power Supply
B.) The output of the power input conditioning stage is

typically a DC voltage source. This DC voltage provides the energy source for the Inverter stage. The Inverter stage converts the DC source to a high frequency AC signal. Many different inverter topologies
exist for power supplies. However, the high voltage
power supply has a few factors which may dictate the
best approach.
Specific performance concerns, technology advances

and application information are presented for the designer,
specifier and user of high voltage power supplies for
analytical instrumentation.
Typically, the Inverter generates a high frequency AC signal which is stepped up by the HV transformer. The reason for the high frequency generation is to provide high
performance operation with reduced size of magnetics
and energy storage capacitors. A problem is created when
a transformer with a high step up ratio is coupled to a high
frequency inverter. The high step up ratio reflects a parasitic capacitance across the primary of the high voltage
transformer. This is reflected at a (Nsec:Npri)2 function.
This large parasitic capacitor which appears across the
primary of the transformer must be isolated from the
Inverter switching devices. If not, abnormally high pulse
currents will be present in the Inverter.
INTRODUCTION
High voltage power supplies are a key component in many

analytical instruments. By the nature of analytical applications, test equipment, methods and data must show
consistent results. The high voltage power supply, being a
critical component within the instrument, must perform
consistently also. The high voltage power supply has
unique concerns which differentiate it from conventional
power supply requirements. By understanding these concerns, the designer and user of Analytical Instrumentation
can gain beneficial knowledge.
Another parameter which is common to high voltage

power supplies is a wide range of load operations. Due to
the presence of high voltage, insulation breakdown, i.e.
tube arcing, is commonplace. The inverter robustness and
control loop characteristics must account for virtually any
combination of open circuit, short circuit and operating
load conditions.
BASIC HIGH VOLTAGE POWER SUPPLY
A.) Figure 1 shows the basic building blocks of most high

voltage power supplies. The Power Input stage provides conditioning of the input power source. The
input power source may have a wide range of input
voltage characteristics. AC sources of 50Hz to 400Hz
at <24V to 480V are common. DC sources ranging
from 5V to 300V can also be found. The power stage
page 18
High Voltage Power Supplies for Analytical Instrumentation
High Voltage Power Supplies for

Analytical Instrumentation
ABSTRACT
SEC.3
18
In addition to wide load variations, virtually all analytical

instruments need to resolve very low signal levels and
contain high gain circuitry. Noise sources, such as power
T E C H N I C A L
R E S O U R C E S
ARTICLES
SEC.3
page 19
supply inverters must be considered. The Inverter can be

a likely source of noise due to the high DI/Dt and DV/Dt
created when the Inverter power devices switch on and
off. The best approach to reduce the noise source is to
have a resonant switching topology. Low output ripple, low
input power source ripple and good shielding practices are
also important.
techniques can be successful. The high voltage output stage also provides feedback and monitoring signals which will be processed by the power supply
control circuits. All of these components are typically
insulated from ground level to prevent arc over. The
insulation materials vary widely, but typical materials
are: air, SF6, insulating oil, solid encapsulants (RTV,
epoxy, etc.). The insulating material selection and
process control may be the most important aspect of
a reliable high voltage design.
All of these concerns, as well as reliability and cost, must

be addressed in the High Voltage Power Supply Inverter
topology.
E.) Control circuits are the glue to keep all of the power
stages working together. Circuit complexity can range
from one analog I.C. to a large number of I.C.s and
even a microprocessor controlling and monitoring all
aspects of the high voltage power. However, the basic
requirement which every control circuit must meet is
to precisely regulate the output voltage and current as
load, input power, and command requirements dictate. This is best accomplished by a feedback control
loop. Figure 3 shows how feedback signals can be
used to regulate the output of the power supply. Conventional regulation of voltage and current can be
achieved by monitoring the output voltage and current
respectively. This is compared to a desired (reference) output signal. The difference (error) between
the feedback and reference will cause a change in the
inverter control device. This will then result in a
change of power delivered to the output circuits.
C.) The High Voltage Transformer is, historically, where

most of the "Black Magic" occurs. In reality, there is
no magic. Complete understanding of magnetics design must be coupled with intense material and
process control. Much of the specific expertise involves managing the high number of secondary turns,
and the high peak secondary voltage. Due to these
two factors, core geometry, insulation methods and
winding techniques are quite different than conventional transformer designs. Some areas of concern
are: volts/turn ratings of the secondary wire, layer to
layer insulating ratings, insulating material dissipation
factor, winding geometry as it is concerned with parasitic secondary capacitance and leakage flux, impregnation of insulating varnish to winding layers, corona
level and virtually all other conventional concerns
such as thermal margins, and overall cost.
D.) The high voltage output stage is responsible for rectification and filtering of the high frequency AC signal
supplied by the high voltage transformer secondary
(Figure 2). This rectification and filtering process in
variably utilizes high voltage diodes and high voltage
capacitors. However, the configuration of the components varies widely. For low power outputs, conventional voltage multipliers are used. For higher power,
modified voltage multipliers and various transformer
Fig. 3
Fig. 2
Typical High Voltage Output Stage
19
Power Supply Control Loops
In addition to the voltage and current regulation, other parameters can be precisely regulated. Controlling output
power is easily accomplished by an X Y = Z function, (V
I = W), and comparing it to the desired output power reference. Indeed, any variable found within Ohm's law can
be regulated, (resistance, voltage, current and power). In
addition, end process parameters can be regulated if they
are effected by the high voltage power supply (i.e. X-ray
output, flow rates, etc.).
T E C H N I C A L
R E S O U R C E S
ARTICLES
SEC.3
page 20
INVERTER TOPOLOGIES
The two approaches have two distinct differences. The

parallel loaded topology more closely resembles a voltage
source, while the series loaded topology resembles a current source. Each have advantages, but typically, the parallel loaded topology is used in low power applications,
and the series loaded topology is used in high power operations. Many reasons exist for this differentiation of use
with power level, but there are a few dictating reasons why
each cannot be used in the others domain. To understand
this we need to visualize the reflected capacitor and what
happens to this capacitor during an output short circuit.
This is of primary importance because under a short circuit condition the parasitic capacitance is reduced by the
reflected secondary load, in this case zero ohms. In the
low power application, the series inductor is of a relatively
high impedance, (due to its VA requirements), and provides
Vt/L current limiting for the inverter switching devices.
As mentioned above, there are a wide variety of Inverter

topologies existing today. However, the choice of Inverter
topologies for a high voltage power supply may be governed by two factors:
Must isolate reflected parasitic capacitance
Must be low noise
Luckily, there is one general approach which meets both

requirements. This approach is resonant power conversion. Resonant topologies utilize a resonant tank circuit for
the generation of the high frequency source. Figures 4 and
5 show two implementations of the resonant approach. Both
successfully isolate the reflected capacitance by a series inductor. In some cases, the reflected capacitance (CR), and
the series inductor (LR) comprise the tank circuit. This is
known as a series resonant/parallel loaded topology. In
other cases, a capacitor is connected in series with the inductor to form a series resonant/series loaded topology.
In the high power, the series inductor is of substantially

lower impedance, and does not provide inherent current
limiting. For this reason, a series loaded circuit is used. It
can be seen by Figure 6, that a series loaded circuit, when
operated outside its resonant tank frequency, resembles a
current source inherently limiting the current capabilities
and thereby protecting the switching devices. (Figure 6)
Fig. 4 Resonant Flyback/Forward Converter

Fig. 6 Series Resonance
Fig. 5 Half Bridge/Full Bridge
20
Still other reasons exist why a series loaded circuit cannot

be used at low power. It can be seen that the series capacitor will support a voltage dictated by the Q of the resonant
circuit and the applied voltage. In all cases, this voltage is
seen across the total circuit capacitance, the series capacitor, and the parasitic capacitor. In the low power application the ratio of the series C to the parallel C is very high
(again due to the VA requirements of the tank). This effectively creates a voltage divider, with most of the voltage
appearing across the series C. This results in a significantly lower voltage applied to the transformer, thereby
limiting high secondary voltages. If higher turns are added,
more reflected capacitance is created and eventually no
additional secondary volts can be generated.
T E C H N I C A L
R E S O U R C E S
ARTICLES
page 21
OUTPUT STABILITY, REGULATION AND REPEATABILITY
The instrument designer may choose to have one vendor

provide all of the power supply requirements. This is very
common in the high voltage area due to the expertise required when dealing with related high voltage circuits (i.e.
filament isolation requirements). For the high voltage
power supply designer this means an expertise in virtually
all aspects of power conversion technology, not just high
voltage power supplies. For example, it is not uncommon
to find filament power supplies providing greater than 100
amps at 20 volts. In addition, this output circuitry may
need isolation as high as 100,000 volts. Even motor control expertise is used in new high voltage technology.
As stated previously, the importance of consistent results

is paramount in the analytical process. The high voltage
power supply must be a source of stable and repeatable
performance. Variations in the output voltage and current
will usually have direct effects on the end results and
therefore must be understood as a source of error. In high
voltage power supplies, the voltage references that are
used to program the desired output can be eliminated as a
source of significant error by the use of highly stable voltage reference I.C.s. Typical specifications of better than
5ppm/C are routine. Similarly, analog I.C.s (op amps,
A/D, D/A's, etc.) can be eliminated as a significant source
of error by careful selection of the devices. [1]
CONCLUSION
There remains one component, unique to high voltage

power supplies, which will be the major source of stability
errors: the high voltage feedback divider. As seen in
Figure 2, the high voltage feedback divider consists of a
resistive divider network. This network will divide the
output voltage to a level low enough to be processed by
the control circuits (i.e. <10vdc).
This paper presented an overview of areas that are specific to the high voltage power supply. The high voltage
power supply has unique concerns which differentiate it
from standard off the shelf products. The designer, specifier and user of high voltage power must be aware of
these concerns, in order to insure the best possible
results. The technological advances in power conversion
are occurring at such rapid rates that is it difficult for an
instrument designer to undertake full responsibility of the
high voltage power supply design. This responsibility,
therefore, must be shared by the supplier of the high
voltage power supply and the instrument designer.
The problem of stability in this network results from the

large resistance of the feedback resistors. Values of >100
megohms are common. (This is to reduce power dissipation in the circuit and reduce the effects of temperature
change due to self heating). The large resistance and the
high voltage rating requires unique technology specific to
high voltage resistors. The unique high voltage resistor
must be "paired" with a low value resistor to insure ratio
tracking under changes of temperature, voltage, humidity
and time.
As discussed in this paper, advanced power conversion

technology, components, materials, and process are required for reliable high voltage design. In addition, safety
aspects of high voltage use requires important attention.
High voltage sources can be lethal. The novice user of
high voltage should be educated on the dangers involved.
A general guideline for safety practices is found in IEEE
standard 510-1983 "Recommended Practices for Safety
in High Voltage and High Power Testing [4]".
In addition, the high value of resistance in the feedback

network means a susceptibility to very low current interference. It can be seen that currents as low as 1 X 10-9 amps
will result in >100ppm errors. Therefore, corona current effects must seriously be considered in the design of the resistor and the resistor feedback network. Also, since much
of the resistor technology is based on a ceramic core or
substrate, piezoelectric effects must also be considered. It
can be demonstrated that vibrating a high voltage power
supply during operation will impose a signal, related to the
vibration frequency, on the output of the power supply.
REFERENCES:
1.)
Precision Monolithics Inc. (PMI), "Analog I.C. Data Book, vol. 10.
3.)
D. Chambers and C. Scapellati, "New High Frequency,

High Voltage Power Supplies for Microwave Heating Applications",
Proceedings of the 29th Microwave Power Symposium, July 1994.
2.)
AUXILIARY OUTPUTS
In many applications of high voltage, additional power

sources are required for the instrument. In many cases,
these auxiliary power sources work in conjunction with the
high voltage power supply. Such examples are: Filament
(heater) power supplies as found in every X-ray tube, bias
(grid) control supplies, focus power supplies, and low voltage power requirements for other related control circuitry.
SEC.3
4.)
21
D. Chambers and C. Scapellati, "How to Specify Today's High

Voltage Power Supplies", Electronic Products Magazine,
March 1994.
IEEE Standard 510-1983, IEEE Recommended Practices for

Safety on High voltage and High Power.
T E C H N I C A L
R E S O U R C E S
ARTICLES
page 22
High Voltage Power Supplies for Electrostatic Applications
High Voltage Power Supplies for

Electrostatic Applications
ABSTRACT
SEC.3
than 24V to 480V are common. DC sources ranging from

5V to 300V can also be found. It is critical for the user to
understand the input voltage requirement as this will impact overall system use and design. Regulatory agencies
such as Underwriters Laboratory, Canadian Standards Association, IEC and others are highly involved with any circuits connected to the power grid. In addition to powering
the main inverter circuits of the power supply, the input
voltage source is also used to power auxiliary control circuits and other ancillary power requirements. The input filter stage provides conditioning of the input voltage source.
by Cliff Scapellati
High voltage power supplies are a key component in electrostatic applications. A variety of industrial and scientific
applications of high voltage power supplies are presented
for the scientist, engineer, specifier and user of electrostatics. Industrial processes, for example, require significant
monitoring of operational conditions in order to maximize
product output, improve quality, and reduce cost. New advances in power supply technology provide higher levels
of monitoring and process control. Scientific experiments
can also be influenced by power supply effects. Contributing effects such as output accuracy, stability, ripple and
regulation are discussed.
INTRODUCTION
The use of high voltage in scientific and industrial applications is commonplace. In particular, electrostatics can be
utilized for a variety of effects. Broadly stated, electrostatics is the study of effects produced by electrical charges or
fields. The applications of electrostatics can be used to
generate motion of a material without physical contact, to
separate materials down to the elemental level, to combine materials to form a homogeneous mixture and other
practical and scientific uses. By definition, the ability of
electrostatic effects to do work requires a difference in
electrical potential between two or more materials. In most
cases, the energy required to force a potential difference
is derived from a high voltage source. This high voltage
source can be a high voltage power supply. Today's high
voltage power supplies are solid state, high frequency designs, which provide performance and control unattainable
only a few years ago. Significant improvements in reliability, stability, control, size reductions, cost and safety have
been achieved. By being made aware of these improvements, the user of high voltage power supplies for electrostatic applications can benefit. Additionally, unique
requirements of high voltage power supplies should be
understood as they can affect the equipment, experiments, process or product they are used in.
Fig. 1 Simplified Schematic Diagram of

a High Voltage Power Supply
This conditioning is usually in the form of rectification and

filtering in ac sources, and additional filtering in dc
sources. Overload protection, EMI, EMC and monitoring
circuits can also be found. The output of the input filter is
typically a dc voltage source. This dc voltage provides the
energy source for the inverter. The inverter stage converts
the dc source to a high frequency ac signal. Many different
inverter topologies exist for power supplies. The high voltage power supply has unique factors which may dictate
the best inverter approach. The inverter generates a high
frequency ac signal which is stepped up by the HV transformer. The reason for the high frequency generation is to
provide high performance operation with reduced size of
magnetics and ripple reduction storage capacitors. A problem is created when a transformer with a high step up
ratio is coupled to a high frequency inverter. The high step
up ratio reflects a parasitic capacitance across the primary
of the high voltage transformer. This is reflected as a
(Nsec:Npri)2 function. This large parasitic capacitor which
appears across the primary of the transformer must be
isolated from the inverter switching devices. If not, abnormally high pulse currents will be present in the inverter.
OPERATIONAL PRINCIPLES OF HV POWER SUPPLIES
The input voltage source may have a wide range of voltage characteristics. AC sources of 50Hz to 400Hz at less
22
Another parameter which is common to high voltage power

supplies is a wide range of load operations. Due to the
presence of high voltage, insulation breakdown is common-
T E C H N I C A L
R E S O U R C E S
ARTICLES
SEC.3
page 23
place. The inverter robustness and control loop characteristics must account for virtually any combination of open circuit, short circuit and operating load conditions. These
concerns as well as reliability and cost, must be addressed
in the High Voltage Power Supply Inverter topology.
The high frequency output of the inverter is applied to the
primary of the high voltage step-up transformer. Proper
high voltage transformer design requires extensive theoretical and practical engineering. Understanding of magnetics design must be applied along with material and
process controls. Much of the specific expertise involves
managing the high number of secondary turns, and the
high secondary voltages. Due to these factors, core geometry, insulation methods and winding techniques are quite
different than conventional transformer designs. Some
areas of concern are: volts/turn ratings of the secondary
wire, layer to layer insulating ratings, insulating material
dissipation factor, winding geometry as it is concerned
with parasitic secondary capacitance and leakage flux, impregnation of insulating varnish to winding layers, corona
level and virtually all other conventional concerns such as
thermal margins, and overall cost.
respectively. This is compared to a desired (reference)

output signal. The difference (error) between the feedback
and reference will cause a change in the inverter control
device. This will then result in a change of power delivered
to the output circuits.
In addition to the voltage and current regulation, other parameters can be precisely regulated. Controlling output
power is easily accomplished by an X Y = Z function,
(V I = W), and comparing it to the desired output power
reference. Indeed, any variable found within Ohm's law
can be regulated, (resistance, voltage, current and power).
In addition, end process parameters can be regulated if
they are effected by the high voltage power supply (i.e.
coatings, flow rates, etc.).
The high voltage multiplier circuits are responsible for rectification and multiplication of the high voltage transformer
secondary voltage. These circuits use high voltage diodes
and capacitors in a "charge pump" voltage doubler connection. As with the high voltage transformer, high voltage
multiplier design requires specific expertise. In addition to
rectification and multiplication, high voltage circuits are
used in the filtering of the output voltage, and in the monitoring of voltage and current for control feedback. Output
impedance may intentionally be added to protect against discharge currents from the power supply storage capacitors.
Fig. 2 Power Supply Control Loops
HIGH VOLTAGE REGULATION
These high voltage components are typically insulated

from ground level to prevent arc over. The insulation materials vary widely, but typical materials are: air, SF6, insulating oil, solid encapsulants (RTV, epoxy, etc.). The insulating
material selection and process control may be the most
important aspect of a reliable high voltage design.
Control circuits keep all of the power stages working together. Circuit complexity can range from one analog I.C.
to a large number of I.C.s and even a microprocessor
controlling and monitoring all aspects of the high voltage
power. However, the basic requirement which every
control circuit must meet is to precisely regulate the output
voltage and current as load, input power, and command
requirements dictate. This is best accomplished by a feedback control loop. Fig. 2 shows how feedback signals can
be used to regulate the output of the power supply.
Conventional regulation of voltage and current can be
achieved by monitoring the output voltage and current
The importance of a regulated source of high voltage

and/or constant current is critical to most applications
involving electrostatics. Variations in output voltage or
current can have direct effects on the end results and,
therefore, must be understood as a source of error. In
high voltage power supplies, the voltage references that
are used to program the desired output can be eliminated
as a source of significant error by the use of highly stable
voltage reference I.C.s. Typical specifications of better
than 5ppm/C are routine. Similarly, analog I.C.s (op
amps, A/D D/A's, etc.). can be eliminated as a significant
source of error by careful selection of the devices.
23
There remains one component, unique to high voltage

power supplies, which will be the major source of stability
errors: the high voltage feedback divider. As seen in Fig.
1, the high voltage feedback divider consists of a resistive
divider network. This network will divide the output voltage
T E C H N I C A L
SEC.3
R E S O U R C E S
ARTICLES
page 24
to a level low enough to be processed by the control circuits. The problem of stability in this network results from
the large resistance of the feedback resistors. Values of
>100 megOhms are common. (This is to reduce power
dissipation in the circuit and reduce the effects of temperature change due to self heating). The large resistance and
the high voltage rating requires unique technology specific
to high voltage resistors. The unique high voltage resistor
must be "paired" with a low value resistor to insure ratio
tracking under changes of temperature, voltage, humidity
and time.
A typical feature that can be implemented into a high voltage power supply is an "ARC Sense" control. Fig. 3 shows
a schematic diagram of an arc sense circuit. Typically, a
current sensing device such as a current transformer or
resistor is inserted in the "low voltage side" of the high
voltage output circuits.
Typically, the arc currents are equal to:
I = (E/R) (1)
where I = Arc current in amperes.
E = Voltage present at high voltage capacitor.
R = Output limiting resistor in ohms.
In addition, the high value of resistance in the feedback

network means a susceptibility to very low current interference. It can be seen that currents as low as 1 X 10-9 amps
will result in >100ppm errors. Therefore, corona current
effects must seriously be considered in the design of the
resistor and the resistor feedback network. Also, since
much of the resistor technology is based on a ceramic
core or substrate, piezoelectric effects must also be
considered. It can be demonstrated that vibrating a high
voltage power supply during operation will impose a
signal, related to the vibration frequency, on the output
The arc current is usually much greater than the normal dc

current rating of the power supply. This is due to keeping
the limiting resistance to a minimum, and thereby the
power dissipation to a minimum. Once the arc event is
sensed, a number of functions can be implemented. "Arc
Quench" is a term which defines the characteristic of an
arc to terminate when the applied voltage is removed. Fig.
4 shown a block diagram of an arc quench feature.
AUXILIARY FUNCTIONS FOR THE HV POWER SUPPLY
In many applications of high voltage, additional control

functions may be required for the instrument. The power
supply designer must be as familiar with the electrostatics
application as the end user. By understanding the application, the power supply designer can incorporate important
functions to benefit the end process.
Fig. 3
Fig. 4 Arc Quench
If shutdown is not desired on the first arc event, a digital

counter can be added as shown in Fig. 5. Shutdown or
quench will occur after a predetermined number of arcs
have been sensed. A reset time must be used so low frequency arc events are not accumulated in the counter. Example: A specification may define an arc shutdown if eight
arcs are sensed within a one minute interval.
Fig. 5
Arc Sense Circuit
24
Arc Count
T E C H N I C A L
R E S O U R C E S
ARTICLES
SEC.3
page 25
A useful application of the arc sense circuit is to maximize

the applied voltage, just below the arcing level. This can
be accomplished by sensing that an arc has occurred and
lowering the voltage a small fraction until arcing ceases.
Voltage can be increased automatically at a slow rate.
(Fig. 6).
Practically stated, as R2 changes impedance there is negligible effect on the current through R1. Therefore, R1 and
R2 have a constant current. In a single power supply
application, this can be accomplished two ways. The first
is to provide an external resistor as the current regulating
device. The second is to electronically regulate the current
using the current feedback control as shown in Fig. 2.
In applications where multiple current sources are
required, it may not be practical to have multiple power
supplies. In this case, multiple resistors can be used to
provide an array of current sources. This is typically used
where large areas need to be processed with the use of
electrostatics. Fig. 8 shows this scheme.
Fig. 6 Automatic Voltage Reduction Circuit
Another feature which can be found in the high voltage

power supply is a highly accurate current monitor circuit.
For generic applications this monitor feature may only be
accurate to milliamperes, or microamperes. However, in
some electrostatic applications accuracy down to femtoamperes may be required. This accuracy can be
provided by the high voltage monitoring circuits. However,
the user of the power supply usually must specify this
requirement before ordering.
Fig. 8
Simple Multiple Current Sources
CONCLUSION
This paper presented information useful to electrostatic

applications using high voltage power supplies. The high
voltage power supply has concerns which differentiate it
from conventional power supplies. The designer of high
voltage power supplies can be a key resource for the user
of electrostatics. Significant control features can be offered by the high voltage power supply. In addition, safety
aspects of high voltage use requires important attention.
High voltage sources can be lethal. The novice user of
high voltage should be educated on the dangers involved.
A general guideline for safety practices is found in IEEE
standard 510-1983 "Recommended Practices for Safety in
High Voltages and High Power Testing [4]".
GENERATING CONSTANT CURRENT SOURCES
In many electrostatic applications, a constant current created by corona effects is desirable. This can be accomplished in a number of unique ways. A constant current
source can be broadly defined as having a source impedance much larger than the load impedance it is supplying.
Schematically it can be shown as in Fig. 7:
REFERENCES:
1.)
C. Scapellati, "High Voltage Power Supplies for Analytical Instrumentation", Pittsburgh Conference, March 1995.
3.)
D. Chambers and C. Scapellati, "New High Frequency, High Voltage Power Supplies for Microwave Heating Applications", Proceedings of the 29th Microwave Power Symposium, July 1994.
2.)
Fig. 7
Practical Current Source
4.)
25
D. Chambers and C. Scapellati , "How to Specify Today's High Voltage Power Supplies", Electronic Products Magazine, March 1994.
IEEE Standard 510-1983, IEEE Recommended Practices for
Safety In High Voltage and High Power Testing.
T E C H N I C A L
R E S O U R C E S
ARTICLES
SEC.3
page 26
Standard Test Procedures for High Voltage Power Supplies
26
T E C H N I C A L
R E S O U R C E S
ARTICLES
SEC.3
page 27
27
T E C H N I C A L
R E S O U R C E S
ARTICLES
SEC.3
page 28
28
T E C H N I C A L
R E S O U R C E S
ARTICLES
SEC.3
page 29
29
T E C H N I C A L
R E S O U R C E S
ARTICLES
SEC.3
page 30
30
T E C H N I C A L
R E S O U R C E S
ARTICLES
SEC.3
page 31
31
T E C H N I C A L
R E S O U R C E S
ARTICLES
SEC.3
page 32
32
T E C H N I C A L
R E S O U R C E S
ARTICLES
SEC.3
page 33
33
T E C H N I C A L
R E S O U R C E S
ARTICLES
SEC.3
page 34
34
T E C H N I C A L
R E S O U R C E S
ARTICLES
SEC.3
page 35
35
T E C H N I C A L
R E S O U R C E S
ARTICLES
SEC.3
page 36
36
T E C H N I C A L
R E S O U R C E S
ARTICLES
page 37
Comparative Testing of Shield Terminations of HV Cables
Comparative Testing of Shield Terminations

of High Voltage Cables
it is the falling resistivity-field characteristic that effectively

pushes the electric field off the shield terminus, where
the field is the strongest.
Alex Pokryvailo, Costel Carp and Cliff Scapellati

Spellman High Voltage Electronics Corporation
Many works were devoted to the field analysis in cable terminations, both analytical and numerical, in linear and
non-linear approaches [1]-[6]. Understandably, they did not
address the space charge formation arising from ionization
around sharp edges. In fact, most designs avoid fields
strong enough to cause ionization. It seems also that no or
little work was dedicated to the investigation of leakage
current (LC) flowing along the cable termination. In lowcurrent, precision HV applications, these currents may be
commensurable with the load current, and being inherently
unstable, can compromise the stability. At the same time,
low-cost design limits the use of high-quality materials
and/or elaborate field control techniques. These limitations
are especially important in open-space connections characterized by very unfavorable stress concentration at the
shield terminus.
Presented at 28th Int. Power Modulators Symp., Las Vegas,

27-31 May, 2008, pp. 576-579.
ABSTRACT
In HV systems, cable terminations are one of the weakest

links. They are especially stressed by the electric field in
free space connections. In this light, several termination
types for polyethylene HV cables were tested for dielectric
strength and leakage current, down to a pico-ampere
level. The tested terminations ranged from simple flush cut
to graded insulation using non-linear insulation materials.
Procedures and results of the testing are described. The
dependencies of leakage current on the applied voltage
for different terminations are presented. Visual patterns of
breakdown are investigated.
In this light, several termination types for polyethylene HV

cables were tested for dielectric strength and LC, down to
a pico-ampere level.
The major results are summarized as follows.
Flush cut shield may have loose strands and presents a danger of the main insulation denting.
Shrink sleeve dominates the ionization phenom
ena, effectively suppressing the corona discharge.
Its influence is much greater at positive polarity of
the shield terminus.
Shield folded back over an O-ring decreases the
electric field, leaves no loose strands and de
creases probability of the main insulation damage.
It can be recommended for DC applications.
Stress grading tapes reduce and greatly stabilize
leakage current at a level of 1nA at 100kV at
room temperature, at positive polarity. They are
less effective in leakage suppression at negative
polarity. They also increase the breakdown voltage
that reaches 130 kV at a 15-cm insulation length,
at both polarities.
EXPERIMENTAL SETUP
Test Rig
The test bench (Fig. 1) comprises a test power supply
(PSU) unit with its HV cable T1, Cable Under Test (CUT)
T2, and measurement and data acquisition means. Two
PSUs (Spellman SL130kV and XRF180kV series) provide
smooth voltage regulation and high stability in the range of
0130kV, 0180kV for positive and negative polarities, respectively. Both HV leads of CUT and that of the PSU
cable are connected together, whereas the CUT shield is
grounded through a current measuring device. A typical
physical implementation is shown in Fig. 2.
INTRODUCTION
In HV systems, cable terminations are one of the weakest

links, whereas the majority of the failures occur at the
ground shield side. This side is especially stressed by the
electric field in free space connections, which is characteristic for some loads. Field control and rigorous technological processes are key to reliable functioning. The first was
realized for a century by stress relief cones in conjunction
with solid dielectric fillings. Later, stress-grading non-linear
materials in form of paint, tapes and tubes were used with
much success (see, e.g., [1]-[3] and their bibliography). In
DC applications, which are the main interest of this paper,
SEC.3
37
Fig. 1. Schematic layout of test rig.
T E C H N I C A L
R E S O U R C E S
ARTICLES
SEC.3
page 38
Test Procedure
For LC, every cable was tested in steps of, typically, 10kV
up to 90kV, and steps of 5kV above 90kV. CUT #6 was not
tested at negative polarity. It is important to note that the
voltage was changed monotonously, ever increasing.
Every test voltage was applied a minute before collecting
the LC data that were transferred from a Keithley picoammeter, model 6487, to a PC using a Keithley ExceLINX
software. A total of 35 counts were taken at each measurement, which lasted 46 s. The results were averaged
over all 35 counts, and the resulting values served for
building the volt-ampere characteristics.
HV leads of all cables were connected physically to the

HV electrode of a voltage divider (Spellman model HVD100) [7] capable of the corona suppressing at the voltage
level up to 130kV as shown by electric field analysis. In
this way, the LCs generated by the ionization (corona)
mechanism at the CUT shield side only are collected and
directed through the picoammeter. In order to exclude the
current originating at the lead end of the CUT, we
screened its shield by a grounded copper electrode.
Fig. 3. CUT #4: left - shield flapped over and held by a SHT; right additional SHT cover. CUT #6 ready for test, SHT 1051727-117 cover
on top of HiK tape.
Fig. 2. Experimental setup.
CUTs
Several CUTs using 2124 Dielectric Sciences Polyethylene (PE) cable were manufactured for testing. All of them
were approximately 2.5-m-long. Their main parameters
are summarized in Table 1, and photos of CUT #4, CUT
#6, as examples, are shown in Fig. 3. Acronyms FC and
SHT stand for Flush Cut and SHrink Tube, respectively
(Alpha irradiated polyolefin SHTs were used). Semiconductive stress grading tape VonRoll 217.21 is SiC-based
and as such exhibits a non-linear behavior. Its conductivity
increases at higher fields effectively suppressing corona.
HiK tape of Dielectric Sciences make is defined as conductive. However, its resistivity is infinite when measured
at low voltage by DVMs. Its datasheet is unavailable.
Table 1. CUTs description.
CUT #3 to CUT #6 were subjected to disruptive voltage

tests. The voltage was raised at a rate of rise of approximately 2kV/s to breakdown, then brought down to a level
by at least 20kV lower that the registered breakdown voltage, and then the test was repeated 2-4 times. In view of
the damage sustained by the shrink insulation and semiconductive tape, we replaced them before testing at the
opposite polarity. No averaging or other statistical processing was applied to the disruptive test data. The flashover
was videotaped to document the flashover pattern. Experimental techniques and measurement means are described further in the body of the text.
RESULTS OF LEAKAGE CURRENT MEASUREMENT
38
For setting a baseline, the first experiment was conducted

with the flush-cut bare cable CUT #1. The LC was stable
in time, especially at positive polarity (grounded shield
negative), and reached 40A at +90kV and 98A at 90 kV. To ensure that the current did originate at the cable
shield, an additional experiment was conducted, in which
the FC was protected by a relatively low-curvature electrode. This brought the current down to less than 3A at
+90kV. Covering the shield termination by SHT suppressed the leakage by orders of magnitude (Fig. 5), especially at positive polarity, which also confirms the current
T E C H N I C A L
R E S O U R C E S
ARTICLES
SEC.3
page 39
origination at the shield cut of the tested termination end.

Reducing the length of the bare PE to l=14.7 cm in CUT
#3 resulted in somewhat higher leakage.
CUT#5 with semiconductive tape (Fig. 7) and CUT #6 with

HiK tape (Fig. 8) had the lowest LCs of the order of 1nA at
100kV at positive polarity. Also, the current was almost
burstless: there were practically no instrument overflows at
200-nA range, whereas other cables could be tested at the
20-A range only at 100kV. CUT #5 was also tested at
negative polarity. The LC was quite large Fig. 9, and even
higher than that of CUT #4 (Fig. 6).
DISRUPTIVE VOLTAGE TESTS
The tests were conducted as indicated in Test Procedure

Section. At positive polarity, CUT #3 had the first flashover
at 104kV along the surface of the test termination. The following breakdowns occurred along the surface of a much
lengthier lead termination at 124kV. The path change may
be attributed to the shield conditioning by the arc trimming
off loose strands.
Fig. 4. CUT#1. Leakage current.
Using an O-ring termination (CUT #4) with SHT, at the

same length of the bared PE, caused the LC drop by an
order of magnitude compared to the flush-cut of CUT #3
(Fig. 6), at both polarities (compare to Fig. 5).
Fig. 7. CUT #5. Leakage current, positive polarity.
Fig. 5. Leakage current, FC, SHT, CUT#2 (l=20cm PE bared length)

and CUT #3 (l=14.7cm) at positive and negative polarity.
Fig. 6. Leakage current of CUT #4 (O-ring, SHT).
Fig. 8. CUT #6. Leakage current, positive polarity.
39
Fig. 9. CUT #5. Leakage current, negative polarity.
T E C H N I C A L
R E S O U R C E S
ARTICLES
SEC.3
page 40
CUT #4 had a different flashover pattern with the spark

bridging the shield and the HV electrode of the HVD-100
through air. First breakdown was at 124.5kV, the consecutive breakdowns occurred at 112kV, 113kV, 117kV. CUT
#5, CUT #6 behaved very similar to each other and distinctly different from the other specimens. They broke
down at 130kV after ~10s exposure. The first flashover
reached the folded end of the shield as indicated by the
arrow in Fig. 10.
ated by a mechanism similar to corona; it also may be

termed as creeping discharge. The onset of tangible currents was around 30kV for both polarities. In agreement
with published data, the corona current is greater at the
positive polarity of the shield, or in the convention of this
report, at negative polarity. A shrink sleeve, besides weakening the field by virtue of electrostatics, leaves place to
but minor ionization in residual air pockets. Charges generated by this mechanism are trapped and neutralize the
external field thus suppressing the discharge and greatly
reducing the LC. A non-monotonous pattern of the curves
is, probably, a result of accumulation and decay of these
charges, a process that may have large time constant in
view of high resistivity of used dielectrics.
Consecutive breakdowns occurred at the same voltage,

but the luminous channel ended at the shrink sleeve end.
At negative polarity, CUT #4 flashed over the PE surface

with the spark anchored at the O-ring. The first breakdown
was at 126kV, the consecutive breakdowns occurred at
the voltages of 109kV, 104kV, almost identical to the case
of the positive polarity. CUT #5 broke down at 136kV after
~5s exposure. The first flashover reached the folded end
of the shield as indicated by the arrow in Fig. 10. The second breakdown occurred at the same voltage, but intense
corona started forming already at 80kV. The rest of the
cables were not tested at negative polarity.
O-ring termination is beneficial for reduction of the external

field from purely electrostatic considerations, although its
advantage over FC is mainly a guarantee of the absence
of loose strands. The O-ring termination was effective at
both polarities.
Stress grading tapes have the effect of pushing the field
away from the shield. At positive polarity, CUT #5, CUT #6
had very stable and low LCs. Their breakdown voltages
were considerably higher then the rest of the designs.
At the opposite polarity, the O-ring termination actually
performed better in terms of leakage. However, the breakdown voltage of CUT #5 was slightly higher than that for
the O-ring termination. In view of only two samples testing, a quantitative comparison may be invalid. Nonetheless, the flashover patterns for these designs are very
indicative. For both polarities, the flashover followed the
short path to the shield with FC and O-ring terminations,
but chose the long path in the case of the semiconductive
and HiK tapes. The latter pattern means that the field at
the shield termination is weakened by the tape, and this
tends to yield higher breakdown voltage.
Fig. 10. CUT#5. Photo of flashover.
ANALYSIS AND DISCUSSION
The electric field distribution in cable terminations is

strongly non-uniform. It deviates from that given by a field
analysis employing constant dielectric and/or conducting
properties by virtue of the presence of space and surface
charges. Actual distribution at DC conditions is greatly influenced by ionization processes: the field is usually reduced by space and surface charges. Semiconductive
tapes act to the same effect, with greater stability and a
benefit of the ionization suppression.
As predicted by the field analysis for FC without SHT, the

field exceeds 100kV/cm and would lead to air ionization.
This is manifested by CUT #1 with exposed shield and erratic short (~1 mm) loose strands protruding from it outwards, which leads to further field enhancement. Large
currents drawn from the shield (Fig. 4) are clearly gener-
It is our opinion that the stress grading tapes are not

necessary for most DC applications but will be a major
enhancement for AC and pulsed applications.
40
As a rule, air gas gaps break down at the same voltage in

repetitive tests, except when electrode conditioning, or
space charge accumulation, or temperature change take
place. In our test, the tendency of lowering the breakdown
voltage values in consecutive tests was quite expressed; it
has a different mechanism. After several flashovers, the
SHTs were punctured and did not suppress LC. It may be
that at negative polarity at a voltage, at which LC reaches
several microamperes, SHT will be damaged in long run
as a result of localized power losses that can be estimated
at a subwatt level.
T E C H N I C A L
R E S O U R C E S
ARTICLES
SEC.3
page 41
The polarity effect on LC for the cables equipped with

SHTs was dramatic: at positive polarity, the LC was by
three orders of magnitude lower than at negative polarity.
We do not have a substantive explanation to this effect.
Numerous publications deal with the influence of dielectric
barrier on the breakdown voltage of gas gaps, with relation
to polarity, barrier placement, kind of gas and its temperature and pressure, etc., but disregard LC. An inference can
be made with reference to the influence of space charge
on the discharge mechanism in strongly non-uniform gas
gaps. On negative polarity (positive shield), negative
space charge attracted to the shield enhances the field,
whereas at positive polarity, the same charge is repelled
and diffused around the shield.
REFERENCES
1.)
2.)
3.)
4.)
5.)
6.)
P. N. Nelson, H.C. Hervig, High Dielectric Constant Materials for

Primary Voltage Cable Terminations, IEEE Trans. on Power Apparatus and Systems, Vol. PAS-103, No. 11, November 1984, pp.
3211-3216.
Wheeler, J.C.G. Gully, A.M. Baker, A.E. Perrot, F.A. Thermal
performance of stress grading systems for converter-fed motors,
IEEE Electrical Insulation Magazine, March-April 2007, Vol. 23,
No. 2, pp. 5-11.
G. C. Stone, E. A. Boulter, I. Culbert, H. Dhirani, Electrical Insula

tion for Rotating Machines, IEEE Press, Wiley, 2004.
S.V. Nikolajevic, N.M. Pekaric-Nad R.M. Dimitrijevic, Optimization
of Cable Terminations, IEEE Transactions on Power Delivery, Vol.
12, No. 2, April 1997, pp. 527-532.
G. Lupb, K Tucci, N. Femia, M. Viielli, Electric Field Calculation in

HV Cable Terminations Employing Heat-Shrinkable Composites
with Non Linear Characteristics, Proc. 4th Int. Conf. on Properties
and Applications of Dielectric Materials, 1994, Brisbane, Australia,
pp. 278-281.
J. Mackevich and J. Hoffman, Insulation Enhancement with HeatShrinkable Components Part 111: Shielded Power Cable, IEEE
Electrical Insulation Mag. July/Aug 1991 -Vo1. 7, N0. 4, pp. 31-40.
http://www.spellmanhv.com/pdf/HVD.pdf
41
T E C H N I C A L
R E S O U R C E S
ARTICLES
Design and Testing of a High-Power Pulsed Load
Design and Testing of a

High-Power Pulsed Load
SEC.3
page 42
long-term resistance stability; this effectively excludes various aqueous solutions, such as copper sulfate aqueous
solutions.
Alex Pokryvailo, Arkady Kogan and Cliff Scapellati

This paper describes the design and testing of a twochannel 52-kW load used in the development of a high
repetition rate capacitor charger.

27-31 May, 2008, pp. 181-184.
ABSTRACT
DESIGN
This paper describes the design and testing of a two-channel 52-kW pulsed load. Its main feature is exceptionally
low parasitic inductance, on the order of 200nH. Such low
inductance was needed in view of microsecond high-current pulses; it was realized by a compact design and careful layout. Small size is a prerequisite for minimizing the
inductance; it was achieved by forced liquid cooling. Noninductive bulk resistors were used at a power rating far
exceeding their specifications detailed for operation in air
and were found adequate for their mission. They were
housed in standard stainless steel drums. The cooling
liquid (water-propylene-glycol mixture) was circulated
through a heat exchanger.
Specifications
The load was designed to the following specifications.
1.) Storage capacitance: C=5.3F (per channel)
2.) Max charge voltage: Vch=1200V
3.) Max Average power: Pav=52kW (26kW per channel)

4.) Pulse width: tpulse5s
5.) Max pulse repetition frequency (PRF): 6kHz
6.) Load inductance:

(per channel, excluding leads) Lload0.2H
7.) Voltage reversal (at maximum charge voltage):

- in normal operation
200V
- in abnormal operation
600V
Multiple aspects of the design are described, including

resistor choice, calculating the load inductance, choice of
busbars, details of kinematic scheme, heat transfer, HV,
safety and other considerations for cooling agents, etc.
Special attention was paid to avoiding turbulent flow that
could result in the resistor cracking. Inductance measurements showed close correspondence with the calculations. High-power testing showed reliable operation with
overheat about 40 K above ambient.
8.) Possibility of reconfiguration to accept pulsed voltage

of several tens of kV.
Circuit ConsiderationsChoice of Resistance
The test circuit can be represented by a capacitor discharge onto r, L circuit, r, L being the load resistance and
inductance, respectively (Fig. 1), the latter including the
leads inductance.
INTRODUCTION
Pulsed resistive dummy loads are widely used in various

HV applications, e.g., testing capacitor charger systems,
nanosecond and picosecond pulsers, etc. Such loads are
characterized by several distinct requirements placing
them apart from more conventional DC or AC loads. One
of the most difficult requirements is providing low parasitic
inductance. It must be of the order of several hundreds of
nH, and tens of nH for microsecond and nanosecond applications, respectively. A natural way of minimizing the
stray inductance is using low-inductive layouts, preferably,
coaxial ones, and minimizing the overall load size. At high
average power and high voltage, the latter is difficult to
satisfy without effective cooling and keeping proper insulation distances. An additional typical requirement is good
Fig. 1. Equivalent circuit for determining load resistance

and inductance.
42
T E C H N I C A L
R E S O U R C E S
ARTICLES
With zero initial conditions, in Mathcad notation, the load

current, i, and the capacitor voltage, v, are given by the
formulae
page 43
Realizing the desired resistance and reconfiguring the

load is convenient with relatively large number of fixed resistors. Their choice is of prime importance influencing the
overall size, cost and reliability. In view of low inductive design, bulk ceramic resistors were chosen. They performed
well in nanosecond applications with forced oil cooling [1],
which was instrumental in obtaining small size, hence low
inductance. Kanthal Globar series 510SP slab resistors
are relatively inexpensive, compact and easy to mount.
The largest parts are specified for the maximum power
dissipation of 150W in air; with oil cooling, based on previous experience, we anticipated good safety margin at a
500-W load. A brief testing of 887SP resistors in static
transformer oil showed that it was capable of bearing the
load of 500-1000W without excessive stress. The main
danger, as indicated by the manufacturer, is bringing the
cooling agent to the boiling point, which would result in the
ceramics cracking. Thus, it is important to avoid turbulent
flow in order to decrease the temperature gradients at the
boundary.
With the target loop inductance L=1.5H, the voltage

reversal of approximately 200 V and tpulse5s are
realized with the load resistance r=0.6 (Fig. 2). A
reversal of 600V can be provided by increasing the
leads inductance to 10H, or decreasing r to 0.25.
Fig. 3 illustrates the capacitor voltage waveforms for
non-inductive discharge (L=0.2H) and artificially
increased L=10H.
Finally, 6.3 20% resistors were chosen. With 48 resistors per channel (~500W per resistor), the connections are
as shown in Fig. 4. The nominal resistance is 0.525, and
the measured value is close to 0.6. The load can be reconfigured to 2.4, 1.2 or 0.3 without major changes.
Fig. 4. Electrical connections (one channel).
Fig. 2. Current and voltage waveforms for L=1.5H; r values

(in SI) as indicated in variables legends. r=1 corresponds to
critically damped discharge.
Fig. 3. Current and voltage waveforms for r=0.6.
SEC.3
Mechanical Layout
The load inductance LLoad is a sum of the resistor assembly inductance and the auxiliary and main busbars inductances. An equivalent circuit (illustrating also the geometrical
arrangement and parasitic resistances) is shown in Fig. 5.
According to it, Lload can be calculated as
43
where LR is the inductance of the resistor pack of 12,

and Laub, Lmb are the auxiliary and main busbars
inductances, respectively.
T E C H N I C A L
R E S O U R C E S
ARTICLES
SEC.3
page 44
Fig. 5. Equivalent circuit of resistive load accounting for

parasitic inductances and coolant conductance.
Minimizing the volume occupied by the magnetic field is

key to achieving low inductance. With this in mind the resistors were grouped twelve in parallel in one plane, the
return path being provided by another group of twelve (see
photo Fig. 6a). The inductance calculation for such an
arrangement may be performed for a flat busbar approximation using the following formula [2]:
Fig. 6. Resistive load being immersed into coolant (one channel).

Load is fully isolated from drum.
Kinematic Diagram
The system works on a closed cycle. The cooling agent is
circulated through the two vessels with loads by means of
a pump and gives heat away in a heatsink provided by a
fan (Fig. 7). The flow is monitored by flowmeters, and the
flow rate can be roughly regulated by valves installed on
the drums. The hosing system is symmetrical with regard
to the loads; no other special means for balancing the load
was designed. Overheat condition that may occur following the pump failure, clots, etc., is prevented by interlocking provided by thermoswitches monitoring the drum
temperatures.
where 0 is the permittivity of free space, d is mean distance between the bars, b, c are the bar thickness and
width, respectively, f, are tabulated values. For the resistor assembly, d=0.06 m, b=0.02 m, c=0.3 m, f=0.8,
=0.002, which yields L=2.5? 10-7 H/m, or
LR=7.5? 10-8 H for the resistor pack having a length of
~0.3 m. This calculation was also verified by finite element
analysis. Since there are two packs connected in parallel,
their inductance is halved (see equivalent circuit Fig. 5).
The auxiliary and main busbars inductances Laub, Lmb
add ~100nH, so the overall load inductance was expected
not to exceed 200300nH. Actual measurement provided
a value of L=200nH (Quadtech 1920 LCR meter, measurement taken at 10kHz).
The resistor assembly fits into a standard 20-gal stainless
steel drum (Fig. 6b) and is suspended by the main busses
on a Lexan lid that serves also as a bushing.
Fig. 7. Kinematic diagram of cooling system.
44
T E C H N I C A L
R E S O U R C E S
ARTICLES
Cooling Agents
Insulating liquids, such as transformer or silicone oil have
good dielectric properties and satisfactory cooling capability, and thus would be an ideal choice. The required flow
rate can be calculated using the formula,
page 45
assembly. The temperature rise may decrease this value

greatly, by an order of magnitude for 2030K, as inferred
from [3], [4].
Analyzing possible load connections Fig. 1, we note that

option b, when the load is tied to ground is preferable in
that the voltage is applied to the coolant only during the
capacitor discharge, and thus the coolant is stressed
during several s only. The parasitic current then flows
between the resistor assemblies (resistances RlbR) and
between the resistors and the drum (resistances RRD)
see Fig. 5. In option a, the voltage across the coolant
resides all the time during the charge, when the current
flows through RRD, and until the capacitor has been
discharged.
where P is the dissipated power, P=52kW=177,000

BTU/hr, cp is specific heat capacity, or just specific heat, at
constant pressure, and T is the target temperature difference. Assuming T=50C between the drum and the outlet of the heat exchanger, we calculate the mass flow rate
Qm per channel for oil with cp=2kJ/kgK Qm0.5kg/s, or
the volumetric flow rate Qv30l/min (8 gal/min). Such
flow rate can be easily provided by conventional pumps.
However, the problem in using oil is poor safety related to
flammability and risk of spillage. Therefore, notwithstanding concerns about dielectric strength and corrosion, we
considered Ethylene Glycol (EG), Propylene Glycol (PG)
and their water mixtures used widely as antifreezes.
Deionized water was discarded in view of expected corrosion and loss of dielectric properties over prolonged service.
We note that in the present implementation our primary

concern resides with the resistance stability, and not with
dielectric strength: the insulation distances are several
centimeters and are ample enough to hold, probably,
hundreds of kV at microsecond durations. We do not have
substantive information on the dielectric properties of
water-glycol mixtures at much longer pulses; however,
some useful estimations can be made to this end. The
power dissipation in the liquid is,
EG and its water mixtures have been used in pulsed

power (see, e.g., [3], [4]), mainly owing to large permittivity
(40 for EG). For withstanding long pulses (several microseconds and longer) water should be clean, and the
solution chilled.
or 1 MW at Vch=1200V and Rliq=1.44 (see Test Results,

following). If applied continuously, such power would bring
the mixture to boiling, which can be considered as coinciding with breakdown at long pulses. Thus, the time to
breakdown can be estimated as
Literary data on resistivity of EG and PG, and especially

their solutions, are difficult to find. The only authoritative
reference to this property was found in [5]. Some additional information is contained in [6]. According to [5], EG
resistivity is 104? m at 20C. A short test was done inhouse to estimate this parameter. Two flat electrodes with
the area of 7cm2, distanced by 0.5 mm, were immersed
into liquid. A Prestone EG-based coolant (presumably,
97% EG) had 140? m at room temperature at a DC
voltage of 10V. Deionized water had 0.7? 104? m at
200V, so it was assumed that the mixture would have resistivity not less than that of EG. Curiously, the measured
values can be considered favorable in the light of experimental data [7], where the maximum of the dielectric
strength for electrolytes, in quasi-uniform fields under the
application of long oblique pulses, was found at
23.5? 102? m.
Obviously, the surrounding liquid acts as a shunt for the

load resistors. For the described geometry, the coolant
shunt resistance (see Fig. 5) may be estimated at 10 at
room temperature, considerably larger than the resistor
SEC.3
assuming adiabatic heating and constant Rliq. For the

liquid mass m=70kg, T=50 K, cp=3.56 kJ/kg? K we
calculate =12s. Such a situation, although hypothetical in
view of the necessity to invest hugely excessive power to
sustain the storage capacitor charged, cautions against
connection Fig. 1a.
45
EG is highly toxic, so eventually a Prestone PG diluted by

deionized water in a proportion of 50%-50% was chosen
as a coolant. PG specific heat of 2.51 kJ/kg? K is close to
that of EG (2.41 kJ/kg? K) [8], and in 50%-50% water mixture cp=3.56 kJ/kg? K, about 85% of specific heat of
water. Thus, the flow rate can be considerably lower than
that for oil circulation.
T E C H N I C A L
R E S O U R C E S
ARTICLES
TEST RESULTS
SEC.3
page 46
ACKNOWLEDGEMENT
Prolonged runs at full power of 52kW showed that the

drums temperature (measured in the midsection using
thermocouples) was 60C70C (depending on ambient
temperature and the position of the heat exchanger) at a
flow rate of 20l/min. The ambient temperature in the test
compartment was maintained by a chiller at 23C, although the temperature around the drums was considerably higher. No sign of resistors degradation except steel
tabs rusting was noted; the coolant, however, became
opaque and slimy, and the busbars were also coated with
slime. The coolant resistance as measured at high current
of up to 3A using a DC power supply varied from 9 at
11C (fresh mixture, kept in the drum for about a month) to
2.8 at 18C (aged mixture), to 1.2 at 54C (aged mixture). This corresponds to the observed increase of the
discharge current by ~10% at hot conditions (67C) compared to cold operation (23Csee Fig. 8).
The authors thank Mr. C. Carp, Mr. R. MacArthur,

Mr. J. LaMountaine and Mr. D. Ryan, all of Spellman
High Voltage, for valuable help in design and conduction
of the experiments.
REFERENCES
1.) A. Pokryvailo, M. Wolf, Y. Yankelevich, S. Wald et al., High-Power

Pulsed Corona for Treatment of Pollutants in Heterogeneous Media,
IEEE Transactions on Plasma Science, Vol. 34, No. 5, October 2006,
pp. 1731-1743.
2.) P. L. Kalantarov and L. A. Zeitlin, Inductance Calculation, 3rd Ed.,

Leningrad, EnergoAtomIzdat, 1986 (in Russian).
D.B. Fenneman and R.J. Gripshover, High Power Dielectric
Properties of Water Mixtures, Proc. 2nd Pulsed Power Conf., 1983,
pp. 302-307.
3.) M. Zahn, Y. Ohki, D. B. Fenneman, R. J. Gripshover, and V. H.

Gehman, Dielectric Properties of Water and Water/Ethylene Glycol
Mixtures for Use in Pulsed Power System Design, Proc. IEEE, vol.
74, No. 9, Sept. 1986, pp. 1182-1221.
Electro-corrosion that is disregarded in short-pulsed systems is an important issue for investigation for this application. However, it is beyond the scope of this paper.
4.) Encyclopedia of Chemistry, vol. 5, p. 984. Ed. N. Zefirov, Bolshaya

Rossijskaya Enziklopedia, Moscow, 1998 (in Russian).
5.) J. Liu, X. Cheng, J. Pu, J. Zhang, Experimental Study of the

Electrical Characteristics of Ethylene Glycol/Water Mixtures in the
Microsecond Regime, IEEE Electrical Insulation Mag., Nov/Dec
2007Vol. 23, No. 6, pp. 20-25.
6.) Impulse Breakdown of Liquids, Ed. V. Y. Ushakov, Springer, 2007,

p. 283-284.
7.) CRC Handbook of Physics and Chemistry, 82nd Ed., Ed. D. R. Lide,
CRC Press, 2002.
Fig. 8. Capacitor voltage and load current at 23 0C.
46
T E C H N I C A L
R E S O U R C E S
ARTICLES
Accurate Measurement of On-State Losses of

Power Semiconductors
Accurate Measurement of On-State

Losses of Power Semi-Conductors
SEC.3
page 47
Three methods are commonly used:

4. Calorimetric method (see, e.g., [1]);
5. Using calibrated heatsinks;
6. Electrical measurements of the device voltage v
and current i, and then finding the losses E by
integrating:
Alex Pokryvailo and Costel Carp

27-31 May, 2008, pp. 374-377.
ABSTRACT
For safe design, the junction temperature should be kept

within the specified range. Three methods are used most
often for determining the power losses:
1. Calorimetric method;
2. Using calibrated heatsinks;
3. Electrical measurements of the device voltage
and current, and finding the losses by integrating these variables.
where T is the period. The power loss is found as , where

f=1/T.
The first method provides accurate and most reliable results, but is difficult to implement, especially in air-cooled
setups. The second method is simpler but inconvenient
for the breadboard setups with ever-changing cooling
schemes. We will discuss in more depth the third method
as most flexible and understandable for electrical engineers.
The paper concentrates on the third method with the emphasis given to the accurate measurement of the on-state
voltage. The techniques of using non-linear dividers with
deep voltage clamping are discussed. Novel circuits allowing faithful measurements of the on-state voltage along
with good timing resolution of the switching transitions are
proposed. Results of circuit simulations are borne out by
extensive testing. Examples of measurement of the onstate voltage of large IGBT modules and free wheeling
diodes (FWD) are presented. The obtained results are applicable for characterizing various power switches, e.g.,
gas discharge devices.
Eq. (1) works out well only if the current and voltage
measurement are correct. In view of a very large dynamic
range of the voltages in the on- and off states, it is difficult
to devise a one-stop setup, although there are recommendations how to circumvent this problem [2]. One needs
high-quality probes and a good scope; this alone does not
guarantee faithful measurements. Ensuring safety is realized with differential probes, at a price of compromising
the measurement accuracy in view of their limited bandwidth and capacitive effects.
In determining the switching losses, good time resolution
is of prime importance, whereas the dynamic range is less
important. For hard switching topologies, these losses
may be estimated using the datasheets. In soft switching
circuits, the conduction losses dominate, and switching
losses may be often neglected. Here the accurate measurement of the on-state voltage comes to the front plan.
The following discussion concentrates on this problem.
INTRODUCTION
For safe design of switch-mode power conversion systems, the junction temperature, , of power semiconductors
should be kept within the specified range. A practical
method of calculating this parameter is using the following
formulae:
Basic technique of narrowing the dynamic range is voltage

clamping using non-linear dividers (see, e.g., [3]). Fig. 1
shows two examples of such dividers. Implementation a
uses N low-voltage diodes connected in series, so when
the applied voltage drops below NVdf, where Vdf is the
diode forward conduction threshold, there is no current
flowing through R1, and the voltage at the scope input
equals HVm. Circuit b functions similarly.
where Tc is the case temperature, is the junction temperature rise over the device case, Q is the component power
loss, and is the thermal resistance, junction to case, specified by the manufacturer. All the indicated temperatures
can be readily measured; determining the power losses,
involves more effort.
47
T E C H N I C A L
SEC.3
R E S O U R C E S
ARTICLES

page 48
Fig. 1. Schematics of basic nonlinear voltage dividers.
Experimental techniques and measurement means are

described further in the body of the text.
Fig. 2. PSpice simulation of circuit Fig. 1b with Zener diodes.

Net aliases in this and following figures show connectivity
(e.g., source V1 is connected to point coil of circuit Fig. 2).
time constant is on the order of a microsecond. This is

larger than typical switching times and commensurable
with the pulsewidth at high conversion frequency. Fig. 2,
Fig. 3 illustrate this statement. The experiments were conducted with a half-bridge quasi-resonant inverter. A Rogowski coil CWT15 [4] was used for monitoring the
components current. Since it is an essentially AC probe,
the current traces are usually biased. In Fig. 3, the bias in
the emitter current, Ie, was removed numerically.
SHORTCOMINGS AND LIMITATIONS OF BASIC CIRCUITS
Circuits Fig. 1 depict idealized, and if realized, the ideal

devices for measurement of low voltages in high dynamic
range. In reality, there are several factors that limit the applicability of these schemes as given in Fig. 1. We skip
here obvious component ratings constraints.
One limitation is the inertia introduced by the time constant

of the measuring circuit, where Cp=Cpr+Cpd is the capacitance of the scope input (including the probe), Cpr, in parallel with the dynamic capacitance of the diodes/Zener
diodes, Cpd. Passive voltage probes have typical capacitance of 10pF, so with R1=10k, the time constant of the
circuit a may be ~10-7 s, i.e., quite small if the diodes capacitance can be neglected. However, the diodes remain
forward-biased for some time after the voltage HVm drops
below the threshold value, since there is no reverse voltage applied to them. This time may be about 1s for
diodes specified for trr =75ns recovery, such as BYM26E,
as show experiments and PSpice simulations. It takes the
diodes ~0.5s to come to a non-conducting state,
because the reverse current is very small and unable
to evacuate the stored charge fast.
IMPROVED PRACTICAL CIRCUITS
The detrimental action of the Zener capacitance can be

rectified using a fast diode connected in series as shown
in Fig. 4 that simulates the actual circuit (except the Zener
diodes were 1N751A, and the diode was MMBD914). Simulations Fig. 4 correspond to the measurements of Fig. 5.
It is seen that the on-state transition is faster and less
noisy compared to Fig. 3. This is important for the loss calculation using (2). We note that a circuit similar to that of
Fig. 4 is described in [3], but the actual waveforms exhibit
slow ~2s transitions, which might be related to the use of
an unsuitable diode.
Fig. 3. Measurement of collectoremitter voltage Vce
of CM300DC-24NFM
Powerex IGBT
using circuit Fig. 2.
TDS 3024B scope is
floating. In this and
further plots, waveform notes carry
scale information
and types of probes
used.
Using signal diodes with trr of the order of a few nanoseconds resolves the stored charge problem as show simulations with 1N4500 diodes having trr=6 ns. However, these
and similar diodes (in experiments, we used MMBD914,
trr=4ns) have significant forward current of tens of A at
tenths of a volt, which translates to a voltage drop across
R1 of the order of 1V. Thus, large number of diodes should
be connected in series to reduce this effect, with some uncertainty remaining.
The capacitance of Zener diodes, on the opposite of the
diodes use, must be accounted for, and in this case, the
48
T E C H N I C A L
R E S O U R C E S
ARTICLES

SEC.3
page 49
Divider Fig. 4 (forward-biased Zener diodes are redundant) is adequate for Vsat measurement of power transistors (and incidentally, many other types of switches, such
as SCRs, GCTs and gas discharge devices), but cannot
be used for the measurement of the forward voltage drop
of free wheeling diodes (FWD) because it swings negative
relative to the HVm point. (Without the cut-off diode, the
divider is universal, but the transition to the on-state is
slow as indicated in Fig. 2, Fig. 3.) In this case, a bridge
formed by fast diodes around a Zener provides a solution
(Fig. 7).
Fig. 4. Blocking Zener diode capacitance using a fast diode.
Circuit excited by source V1 Fig. 2.
Although measurements Fig. 5 can be believed to be true

in the sense that the voltage between the measurement
points was recorded faithfully, the actual Vce voltage is different from it owing to the IGBT internal inductance LIGBT.
The inductive voltage drop can be deducted from the
measured voltage; a corrected waveform calculated for
LIGBT=20nH is shown in Fig. 6.
Fig. 7. Bridge formed by fast diodes around a Zener diode works

equally well for measurement of positive and negative low voltages
in wide dynamic range. Circuit excited by source V1 Fig. 2.
Fig. 8 shows the trace of an IXYS DSEI 2x61 FWD current

(one module contains two diodes connected in parallel)
together with the voltage trace taken with the divider Fig. 4
(fast diode removed) with the scope floating. The voltage
trace has almost a sine wave form with a slow falltime,
which is a measurement error caused by the inherent
defect of this circuit (Zener diode capacitance).
Using a divider Fig. 7 provides a different picture and is
believed to improve the measurement considerably as
seen in Fig. 9 that shows also an adjusted waveform and
loss curves. Again, the actual forward drop is lower by the
inductive component.
Fig. 5. Measurement of saturation voltage Vsat

(collector-emitter voltage Vce,) of CM300DC-24NFM using circuit
Fig. 4. Scope is floating.
Fig. 6. Vce adjusted for inductive voltage drop (numerical filtering

has been applied). It corresponds to CM300DC-24NFM datasheet.
49
Fig. 8. Trace 2 - Forward drop of FWD IXYS DSEI 2x 61 (negative

part). Clamped positive voltage (diode non-conducting) is off-scale.
Zener diode capacitance (divider Fig. 4) affects the voltage fall time.
T E C H N I C A L
R E S O U R C E S
ARTICLES

SEC.3
page 50
They are less trustworthy in our opinion than their floating counterparts Fig. 5, Fig. 9 (see also the superposition
of the differential and floating measurements Fig. 12),
which can be explained by the probe limited bandwidth
(25MHz for P5200 compared to 500MHz for P6139A),
leads capacitance to ground in addition to a 7pF capacitance of each input (estimated 30pF total), and by the
large voltage swings (~360V at a rail voltage of 600V) of
the inputs relative to ground. Therefore, battery-fed
scopes, such as Tektronix TPS series are preferential for
this task. Even better, universal, and less expensive solution is using regular scopes fed from an uninterruptible
power supply disconnected from mains. Usual safety
precautions should be taken in floating measurements.
Fig. 9. FWD IXYS DSEI 2x61 losses. Plot a green trace is measured
signal; brown trace is Vfwd adjusted for inductive drop LdIfwd/dt
(diode assembly inductance assessed at 5nH). Green and brown
curves plot b match their counterparts in plot a.
Divider Fig. 7, Floating scope.
FLOATING OR DIFFERENTIAL MEASUREMENTS?
SAFETY ISSUES
As a rule, the scope chassis is grounded for safety, and
floating measurements are performed with differential
probes as recommended by scope vendors (see, e.g., [2]).
Our experience shows, however, that the quality is severely compromised compared to the case when the
scope is floating together with the reference point, e.g., the
transistor emitter or the FWD anode. Examples of using a
differential probe P5200 for Vce and FWD forward drop
measurement are shown in Fig. 10, Fig. 11, respectively.
Fig. 11. Trace 3 - Forward drop of FWD IXYS DSEI 2x 61,
two modules in parallel. a high-bandwidth P6139A probe,
b - differential probe.
Both measurements taken with floating scope.
Fig. 10. Differential measurement of Vsat (trace 3 Vce) of CM300DC24NFM Powerex IGBT using circuit Fig. 4. Trace 3 may have some
offset, likely zero is shown by dashed line.
50
T E C H N I C A L
R E S O U R C E S
ARTICLES

Fig. 12. Superposition of differential and

floating measurements of Fig. 11a, b.
CONCLUSION
Divider Fig. 4 is recommended for the measurement of the

on-state voltage of large power switches. Clamping voltage should be adjusted to the expected on-state value
using proper number of zener diodes. Floating measurements provide better accuracy, however, safety rules
should be strictly observed.
ACKNOWLEDGMENT
The authors acknowledge the support of this work given

by Spellman High Voltage Electronics Corporation.
REFERENCES
1.) C. Huang, P. Melcher, G. Ferguson and R. Ness,

IGBT and Diode Loss Measurements in Pulsed Power Operating
Conditions,
Proc. Power Modulator Symposium, 2004, pp. 170-173.
2.) S. Gupta, Power Measurements and Analysis:
Challenges and Solutions, Tektronix White Paper.
3.) A. Calmels, VDS(on), VCE(sat) Measurement in a High Voltage,

High Frequency System, Advanced Power Technology,
Application note APT0407, November 2004.
4.) http://www.pemuk.com/pdf/cwt_mini_0605.pdf
51
SEC.3
page 51
T E C H N I C A L
R E S O U R C E S
ARTICLES
Highly Efficient Switch-Mode 100kV, 100kW

Power Supply for ESP Applications

page 52
cilitating better collection efficiency. A good overview is

provided by [1], [2]. It was noted that Alstom and NWL lead
the market with hundreds of fielded units. Between other
developments, work of Applied Plasma Physics [3],
Genvolt [4], VEI [5] should be mentioned.

High conversion frequency, typically 20-25kHz facilitates

the size reduction. As noted in [2], the HV transformer of
the Alstom SIR weighs about 22 lb, or 1/15 of that for a
60Hz power supply. Other passive components are shrunk
respectively.
Presented at 11th Int. Conf. on Electrostatic Precipitation, Hangzhou,

21-24 Oct., 2008, pp. 284-288.
ABSTRACT
For nearly a century, electrostatic precipitators (ESP) were

driven by line-frequency transformer-rectifier sets. The last
decade has been marked by steady penetration of highfrequency HV power supplies (HVPS) that offer considerable benefits for the industry.
Heat management is one of the main issues for reliability.

It is done by air-cooling (NWL) or liquid cooling (Alstom). It
should be noted that air-cooling schemes seem to be preferential in this industry. In order to realize high efficiency,
almost universally, the converter part of the above HVPS
makes use of series resonance to avoid switching losses.
The theory and practice of such converters is known well
[6], [7]. A natural way for the voltage/current adjustment in
such converters is frequency regulation. Audio noise is not
an issue for the ESP and similar applications.
This paper describes a novel concept and physical

demonstration of an ultra-high efficiency, small size and
low cost HVPS specifically designed for ESP and similar
markets. Key technology includes a modular HV converter
with energy dosing inverters, which operate at above
50kHz with and have demonstrated an efficiency of 97.5%
in a wide range of operating conditions. The inverters output voltages are phase-shifted, which yields an exceptionally low ripple of 1% and a slew rate of 3kV/s combined
with low stored energy. Modular construction allows easy
tailoring of HVPS for specific needs. Owing to high efficiency, small size is achieved without turning to liquid cooling. Controls provide standard operating features and
advanced digital processing capabilities, along with easiness of accommodating application-specific requirements.
This paper describes a novel concept and physical demonstration of an ultra-high efficiency, small size and low cost
HVPS specifically designed for ESP and similar markets.
MAIN SPECIFICATIONS
1.)
Average output power 100kW in the output voltage

range of 90-100kV; derated at lower voltage
3.)
Dynamic Response: slew rate 100kV/ms min (5% to

9 5% of preset voltage). Typically 300kV/ms
2.)
HVPS design and testing are detailed. Experimental current and voltage waveforms indicate virtually lossless
switching for widely-varying load in the full range of the
line input voltages, and fair agreement with simulations.
Calorimetric measurement of losses indicates to a >98.5%
efficiency of the HV section. The overall efficiency is 95%
at full load and greater than 90% at 20% load, with power
factor typically greater than 93%.
4.)
5.)
6.)
7.)
KEYWORDS
Electrostatic Precipitator, ESP Power Supplies, High-Frequency Power Supplies, voltage multiplier
8.)
9.)
INTRODUCTION
For nearly a century, ESPs were driven by line-frequency

transformer-rectifier sets. The last decade has been
marked by a steady penetration of high-frequency HV
power supplies (HVPS) that offer considerable benefits for
the industry: small size, low ripple, fast response, etc., fa-
SEC.3
High frequency ripple component: 1% typically at

100kV, full power.
Output Stored Energy: < 10 J.
Conversion frequency 50kHz
Input Voltage: Three Phase 400VAC +10%, -14%

Power Efficiency: typically > 95% at full power at
100 kV, > 90% at 20kW.
Power factor: > 93% at full power at 100kV, > 75%

at 20kW.
SPARK/ARC WITHSTAND
10.) Overall weight 250kg TBD; HV unit 109kg (240 lbs);

Oil volume less than 60 liter
52
T E C H N I C A L
R E S O U R C E S
ARTICLES

KEY TECHNOLOGY
SEC.3
page 53
The maximum frequency, at which the operation is possible with zero-current crossing (ZCC), in a normalized form,
is given by the equation
The HVPS is built around a modular HV converter (Figure

1). All converter modules M1-MN are fed from a common
Input Rectifier (IR). The modules comprise inverter INV1INVN feeding HV transformers T1-TN that feed voltage
multipliers R1-RN, which voltages are summed by their
DC outputs. Such topology may be termed as inductive
adder. For the 100kV, 100kW rating N=4. Each module is
built for 25kV, 25kW average power and must have high
potential insulation of the secondary winding of the transformer rated at 325kV=75kVDC. This insulation must also
withstand transient voltages arising during the HVPS turnon and turn-off. The number of such transients is determined by the HVPS operating scenario, and mainly by the
sparking rate.
where E is the rail voltage, and both the rail voltage and
the load voltage Vl are referenced to the same side of the
transformer. The conversion frequency f is normalized to
the resonant frequency f0 of the loop formed by the leakage inductance and resonant capacitors:
The topology Figure 1 was investigated long ago. It allows

reduction both of the number of the multiplier stages and
the voltage rating of the HV transformer. The first improves
the compression ratio and reduces drastically the stored
energy. Phase shift of the inverters outputs voltages results in the decrease of the output ripple and in additional
reduction of the stored energy. In this approach, the development costs and time are driven down noting that once a
single module has been developed (including its main insulation), the whole system is realized by a simple combination of the desired number of modules. The penalty is
larger part count and the necessity of high-potential insulation that is not required in conventional Cockroft-Walton
multipliers. However, this insulation is subjected mainly to
DC stresses and therefore ages much slower compared to
an AC stress.
A sample plot of this equation is shown in Figure 3. It

should be noted that the real conversion frequency is
somewhat lower to allow a deadtime of ~1.5s.
The converter cells are centered around half-bridge energy dosing quasi-resonant inverters (Figure 2) [10], [11],
[12]. The principle and theory of operation were put forward in [11]. In normal mode, one of the divider capacitors,
Cdiv, is charged to the rail voltage. When the corresponding switch closes, it discharges through the primary, while
it counterpart recharges to the rail voltage. If the current
path contains an inductance, a sine waveform is generated, and ideally, all the energy stored in Cdiv would be
transferred to the secondary side. If Cdiv discharges fully,
and the current does not fall to zero, the free-wheeling
diodes (FWD) across the capacitors clamp the current preventing the voltage reversal. Thus, the remainder of the
energy stored in the circuit inductance is transferred to the
output (see also Figure 4). The benefits of this topology
are tight control of the energy transfer and inherent limitation of the short circuit current and voltages across the
converter components.
Fig. 1. HVPS block-diagram.
53
T E C H N I C A L
R E S O U R C E S
ARTICLES

SEC.3
page 54
Fig. 2. Inverter with energy dosing capacitors
The inverters operate at approximately 50kHz at full load

with virtually zero switching losses. The leakage inductance of the HV transformers is fully incorporated into the
resonant tank circuits, so no external inductors are necessary. Besides lowering the part count and cost, this feature
is highly beneficial for the chosen multicell resonant topology, since leakage inductance is well repeatable from
sample to sample and does not depend on temperature.
Controls provide standard operating features and advanced digital processing capabilities, along with the easiness of accommodating application-specific requirements.
The output regulation is accomplished by the frequency
control.
Fig. 3. ZCC curves for low (460V), high (592 V) and nominal (525V)
DC rail voltages. Vlnom is nominal load voltage.
Special attention was paid to the determination of the HV

transformer and multiplier losses. This was key to the design of the HV tank. With this purpose, calorimetric measurements of the losses were performed. They yielded a
figure of 344W, with 175W attributed to the transformer
losses, and the rest to the multiplier losses. Thus, the efficiency of the HV section was expected to be >98.5%.
Accounting also for the inverter losses, the converter efficiency was estimated at 97.5%, so the overall efficiency of
95% of the whole HVPS was projected. In view of the
expected high efficiency, it was decided to adopt an aircooling scheme.
EXPERIMENTAL
Single module
Typical waveforms shown in Figure 4 (taken at nominal
line) indicate good resonant switching with no shootthrough currents in the full range of the line input voltages,
and fair agreement with PSpice simulations. The primary
winding was divided into two sections connected in parallel, each commutated by a transistor set, hence the notation halved in the figure caption. The dashed line shows
the start of the FWD conduction. At low line, the FWDs do
not conduct, and the converter operates in a boundary
mode given by (*). These measurements were conducted
with the Powerex IGBTs CM300DC-24NFM. The power
losses were assessed at 50W per transistor (four transistors, or 800W per converter module), and the heat was
easily evacuated using air-cooled heatsinks with overheat
above ambient of less than 40C. The methods of power
loss measurement are detailed in [13].
54
Fig. 4. Nominal line. P=28.7kW. trace 1 primary winding

current (halved); trace 3 collector current (halved);
trace 4 voltage across resonant capacitors. FWD conducts
to the right of dotted line.
T E C H N I C A L
R E S O U R C E S
ARTICLES

SEC.3
page 55
Fig. 7. Same as in Figure 6 at 100 kV@50 kW.

Low line 400 VAC-14 % (345 VAC).
Since the full-wave rectification scheme is used, the phase

shift is /4. PSpice calculations predict 0.223% output voltage ripple peak-to-peak (p-p) with the HVPS shock capacitance of <2nF (Figure 8) at the worst case of high line; the
measured ripple is roughly four times larger, and has a
lower frequency fundamental component (Figure 9), which
can be attributed to the asymmetry of the gate signals,
unequal parasitic capacitances, spread in winding data,
etc. Similar effect was observed in [9]. These simulations
provide also a value of the Power Factor (PF) of 0.943,
which is close to the experimental results.
Fig. 5. Laboratory HVPS.
HVPS Tests
A laboratory HVPS was assembled on a cart as shown in
Figure 5. It comprises three main units: a circuit breaker
protected line rectifier, an inverter section and an oil-filled
HV tank. We note that in this work, the emphasis was on
the converter part; the line rectifier was not optimized.
The HVPS was extensively tested with resistive loads. Figure 6 and Figure 7 show typical phase-shifted primary windings currents (halved) for 100kW and 50kW operation,
respectively. The oscillations after the main current surge
are generated by the resonance between the leakage inductance and parasitic capacitance of the transformers.
Note the absence of the backswing current pulse characteristic for the series resonant schemes under light load.
Fig. 8. HVPS circuit simulation. High line 580V. ripple 0.223% p-p.
PF=0.943. Experimental PF= 0.946 (see Figure 11).
Fig. 6. /4Sphase-shifted primary windings currents (halved) at
100kV@100kW. Nominal line voltage 400 VAC.
55
T E C H N I C A L
SEC.3
R E S O U R C E S
ARTICLES
page 56

gle module, the overall efficiency is 95% at full load and

greater than 90% at 20% load. The power factor was also
satisfactorily high (compare to the simulation Figure 8). At
high and low line, the measurements yielded very similar
results. At higher resistance load, the efficiency and PF
also stayed high (Figure 12).
At the time of writing this paper, long-term runs at 100kV

have been performed up to a power level of 75kW. Fullpower tests were limited to ~40 min. They showed conservative overheat of the major HVPS components. For the
nominal line, the results are summarized in Table 1.
load power, kW
Fig. 9. Ripple at 100 kV across 100Sk load is 0.762 % p-p.
75
100
The dynamic response of the HVPS is exceptionally fast:

the risetime from zero to full output voltage is typically less
than 250s (Figure 10), depending on the line voltage.
With fair accuracy, the dynamic characteristics can be
analyzed using the equation
transistor
baseplate
20
25
FWD
baseplate
18
23
HV tank
27
N/A
Table 1. Overheat of major HVPS components, C.
where all the variables and parameters are reflected to the

same side of the transformer; Cs is the overall capacitance
of the module multiplier. If the frequency is varied during the
charge, PSpice simulations provide much better accuracy.
Fast response is beneficial not only for ESP but medical
applications as well. We note that the risetime practically
does not depend on the load, since the load current is by
an order of magnitude smaller than the current charging
the multiplier capacitors.
Fig. 11. Apparent, Pinapp, and active input power, Pinact, load
power, Pl, efficiency and PF at nominal line for 100k load.
Fig. 10. Risetime across 95k load at nominal line. Trace 2 load
voltage, 20kV/div; trace 1 primary current (halved), 100 A/div.
Figure 11 presents experimental data on the power measurements obtained at nominal line. In accordance with the
simulations and information derived from the work with sin-
56
Fig.12. Same as in Figure 11 for 200k load.
T E C H N I C A L
R E S O U R C E S
ARTICLES

ACKNOWLEDGMENTS
The authors thank their colleagues at Spellman for

massive support of this work, and especially Mr. A.
Lipovich for his contribution to the mechanical design,
and Mr. A. Silverberg for the realization of the phaseshift algorithm.
REFERENCES
1.) K. Parker, Electrical Operation of Electrostatic Precipitators, IEE,

London, 2003, 270pp.
2.) Advanced Electrostatic Precipitator (ESP) Power Supplies Update:

The State-of-the-Art of High-Frequency Power Supplies. EPRI, Palo
Alto, CA: 2006. 1010361.
3.) M. K. Kazimierczuk, D. Czarkowski, Resonant Power Converters,
Wiley, NY, 1995.
4.) R. Erickson and D. Maksimovic, Fundamentals of Power Electronics (Second Edition), Springer, NY, 2001, 912pp.
5.) US Patent 4,137,039, X-Ray Diagnostic Generator, Feb. 23, 1982.

6.) Yu. Petrov and A. Pokryvailo, HV DC-to-DC Converter, Pribory i
Teckhnika Experimenta, v.2, pp. 141-143, 1986, Translation to English Plenum Publishing Corp.
7.) B.D. Bedford and R.G. Hoft, Principles of Inverter Circuits, Wiley,
NY, 1964.
8.) B. Kurchik, A. Pokryvailo and A. Schwarz, HV Converter for Capacitor Charging, Pribory i Tekhnika Experimenta, No. 4, pp.121-124,
1990, Translation to English Plenum Publishing Corp.
9.) M. Wolf and A. Pokryvailo, High Voltage Resonant Modular Capacitor Charger Systems with Energy Dosage, Proc. 15th IEEE Int.
Conf. on Pulsed Power, Monterey CA, 13-17 June, 2005, pp. 10291032.
10.) A. Pokryvailo and C. Carp, Accurate Measurement of on-State

Losses of Power Semiconductors, 28th Int. Power Modulators
Symp., Las Vegas, 27-31 May, 2008.
57
SEC.3
page 57
T E C H N I C A L
R E S O U R C E S
ARTICLES
High Power, High Efficiency, Low Cost Capacitor Charger

Concept and Demonstration
High Power, High Efficiency, Low Cost

Capacitor Charger Concept and Demonstration
SEC.3
page 58
MAIN SPECIFICATIONS

Presented at 17th IEEE International Pulsed Power Conference

June 29-July 2, 2009 Washington DC
ABSTRACT
A 20kJ/s, 10kV, 1kHz repetition-rate technology demonstrator design and testing are described. The goal of the
development was combining high performance and versatility with low-cost design and good manufacturability. This
goal was met using an energy-dosing converter topology
with smart controls adapting the switching frequency in
such a way as to ensure zero-current switching for all
possible scenarios, keeping maximum duty cycle for high
power. The switching is accomplished at a frequency of up
to 55kHz employing relatively slow IGBTs with low conduction losses. High efficiency allows all-air cooled design
that fits into a 19x10x24 rack.
Design guidelines are reviewed. Comprehensive PSpice

models accounting for numerous parasitic parameters and
mimicking controls for the frequency variation were developed, and simulation results are presented. Together with
analytical tools, they predicted a pulse-to-pulse repeatability (PPR) of 0.15%; the measured figures are 0.4% and
0.5% for short- and long-term operation, respectively, at
peak charging and repetition rate. Repeatability analysis is
briefed upon here, and to larger extent, in an accompanying paper. Test methods are described. Typical current and
voltage traces and results of thermal runs are presented.
DESIGN
A charger block-diagram is shown in Figure 1. The

charger comprises a 3-phase input rectifier with soft start
and a smoothing filter, a converter module (CM), an HV divider and control means. Triggered by an external source,
the charger charges capacitor Cs that is discharged onto a
dummy load via a high-power switch DSw.
CM comprises an inverter INV, HV transformer using popular U100/57/25 ferrites, a rectifier R and control means.
The CMs heart is a half-bridge quasi-resonant inverter
with energy dosing capacitors (Figure 2) [1]-[3]. Work [2]
provides the principle and theory of operation. The benefits of this topology are tight control of the energy transfer
and inherent limitation of the short circuit current and voltages across the converter components.
INTRODUCTION
Between numerous capacitor charging applications, a

combination of high voltage, high charging rate (tens of
kJ/s and higher), high pulse repetition rate (PRR), compactness, high efficiency and good pulse-to-pulse repeatability (PPR) is a serious technological challenge. Putting
constraints of low-cost and good manufacturability makes
the charger development even more difficult. They restrict
use of costly switches, e.g., SiC, exotic cooling schemes
and materials, leaving freedom to choose proper circuit
topology and control strategy to increase the switching frequency with the purpose of shrinking the size and improving PPR. This paper describes an attempt to satisfy the
above contradicting requirements within the constraints of
low-cost proven technology.
The maximum conversion frequency is 55kHz at low rail

voltage. The parasitics of the HV transformer together with
capacitors Cdiv form the resonant tank circuit. Standard
components and subassemblies field-tested in thousands
of Spellman HVPS were used throughout for low cost and
reliability.
58
T E C H N I C A L
R E S O U R C E S
ARTICLES

Fig. 1.
SEC.3
page 59
Charger block-diagram.
Fig. 3. Charger a front view; b HV section.
design and interpretation of the experimental data. A

sample of simulated waveforms is given in Figure 4 for the
cases of low- and high line voltage. In these simulations,
Cs was 200 nF, approximately half of that used in the experiments. It is seen that at any moment (except the first
pulse) during the charging cycle ZCS is attained. This was
confirmed experimentally.
Fig. 2. Halfbridge inverter with energy dosing capacitors.
The FPGA-based controls are characterized by their flexibility ensuing from digital processing capabilities. The
standard features include multiple protections (short
circuit, overheat, overcurrent and overvoltage, etc.) and
means of voltage and current setting. Via firmware, an
algorithm is implemented that adapts the switching frequency in such a way as to ensure zero-current switching
(ZCS) for all possible scenarios, keeping maximum duty
cycle for high power. Thus, the switching losses are virtually non-existent, which allows using relatively slow lowcost switches both on the primary and secondary side.
A precision feedback divider was designed for high-fidelity

measurements necessary for good PPR. A risetime of less
than 1s and low temperature drift were realized.
The packaging was made in a 19 rack-mounted chassis,

10 tall, 24 deep. On the front view (Figure 3a), the front
panel borrowed from the ubiquitous SR6 series [4] is seen.
The filling factor is low as shown in Figure 3b, so the unit,
weighing in at 41kg, is relatively light.
Comprehensive PSpice models accounting for numerous
parasitic parameters and mimicking controls for the
frequency variation were developed assisting in both the
59
Fig. 4. PSpice simulation for 460V and 590VDC rail voltage.

Cs=200nF.
T E C H N I C A L
R E S O U R C E S
ARTICLES

EXPERIMENTAL
SEC.3
page 60
Measurement Means
For the measurement of the high-frequency current of the
inverter components, Rogowski probes of PEM make,
model CWT15, were used. The Cs voltage was measured
by a Tektronix P6015A probe. Floating voltage measurements were performed by a differential Tektronix probe
P5200. Efficiency and power factor were measured with a
Voltech power meter, model PM300.
Waveforms
One of the main goals of this work was realizing as high
efficiency as possible by enforcing lossless switching in all
possible scenarios at all charge levels and repetition rates.
The noise immunity of the control circuitry in this sense is
also an important issue. A thorough experimental investigation side by side with PSpice modeling was performed.
We found that no under circumstances ZCS was disturbed. Several screens below illustrate the results. Figure
6a shows Vc and primary winding current I1 in burst operation at a PRR of 1400Hz for Cs=420nF (charge rate of
29.4kJ/s), with the collector current, Ic, of one of the
transistors displayed on expanded scale in Figure 6b.
Fig. 6. Typical waveforms at 10 kV@1000 Hz at low (a)
and nominal (b) line.
At low line (longest charge), Cs=420nF is charged in

750s (Figure 6a), so continuous operation with such load
is limited to a PRR of 1kHz, if ample dead time is desirable
between the shots. At higher line voltage, the charge is
accomplished faster (Figure 5a, Figure 6b). As vividly seen
in Figure 5b, the conversion frequency adapts to keep
high duty cycle yet maintaining ZCS; there are no shootthrough currents. The highest conversion frequency is
55kHz at low line, with very large margin guaranteeing
ZCS even at abnormal line sags.
Repeatability
PRR is an important parameter in capacitor charging applications. It influences stability of various physical
processes ranging from lasing to pulsed X-rays to plasma
chemistry applications. PPR,R, is defined here1 as
Fig. 5. Typical waveforms at highline. PRR=1400Hz in burst,

charge time is 507s. a load capacitor voltage and
primary winding current; b collector current.
60
where VCmax, VCmin and VCavg are maximum, minimum and

average values of the voltage across the storage cap for a
T E C H N I C A L
R E S O U R C E S
ARTICLES

predefined number of pulses. Pulse-to-pulse variability

evolves from several factors:
SEC.3
page 61
The scope was triggered by the EOC event. Note that the
discharge switch DSw is fired on in 20s after EOC. The
first 800 shots were collected with a 500pnt resolution on a
4s/div scale. The waveforms were saved as screen captures, and 80 frames, starting from 121st frame, were
saved in the csv format. An Excel spreadsheet was designed, in which 79 shots2 were processed; they are
graphed in Figure 7 for several rail voltages showing
pulse-to-pulse Vc variation.
1.) Converter remnant energy, Erem, at the End-ofCharge (EOC). This energy can be stored in the
HV transformer magnetic system, its parasitic
capacitance, resonant capacitors, buswork, etc.
Erem may flow wholly or partially to the storage capacitor, so the output voltage will be higher than the
programmed value.
2.) Error in generating EOC signal. This may be caused

by poor-quality feedback, noise, unstable reference
voltage, etc.
3.) Delay, td, between EOC and actual IGBT turn-off. It
comprises digital delays, optocouplers delay, and
IGBT turn-off delay. Even constant td, if commensurable with half-period, affects PPR. Depending on
the circuitry and the components, td can be fractions
of a microsecond, i.e., td is commensurable with
half-period.
We will distinguish here between short-term and long-term

PPR. The former is defined as that derived from N consecutive pulses. In our measurements, N=80, sampled from
121st to the 200th pulse. Thus, short-term PPR is not influenced by thermal drifts, aging of components, etc. It is
affected by the rail voltage variations to the extent of the
high-frequency rail voltage ringing, excluding slow input
changes. Long-term PPR is also influenced by the rail voltage variation in the full defined range, for instance, from
460VDC to 590VDC (corresponding to 400VAC +10%,
-14%). In this report, the reference to long-term PPR is
made in the light of such variations, other parameters
being not controlled.
PPR measurements were taken using the FastFrame capability of a DPO7054 scope. Up to four signals were
monitored simultaneously. The load voltage, Vc, was
measured again by the P6015A probe, but on a 100mV
scale with a 10V offset allowing the signal at end-ofcharge (EOC) fit the screen. In addition, the feedback voltage, Vfdbk (with the same sensitivity and offset), and
primary current were monitored. The shortcoming of these
direct measurements is their low resolution, of the order of
several bits of the scope vertical resolution. Arguably, a
better technique is differential measurement, e.g., monitoring the difference between the feedback voltage and the
programming voltage. In such a way, at EOC the scope
would see virtually zero voltage. In the differential measurement, the feedback voltage was biased with a voltage
equal to the programming value. After finding fair matching
of the Vc and differential Vfdbk data, we continued with direct Vc measurement only.
Fig. 7. Shot-to-shot variability taken with FastFrame.

Cs=420nF, 1kHz reprate, 2kV, 6kV and 10kV settings.
See inset annotations for rail voltage.
1PPR is defined by most vendors as xx%, so 1% in our
measurements corresponds to 0.5% in their definition
Values shown are averages of 50 points, starting from 250pnt of

the acquisition (approximately, the middle part of the screen Figure 8).
61
T E C H N I C A L
R E S O U R C E S
ARTICLES

Three typical screenshots of the overlays of 80 frames are

shown in Figure 8. They correspond to data Figure 7 and
show quite vividly wherefrom the variability, at least partially, evolves. At EOC, the primary current is chopped at
random. If there is a certain pattern (as seen at 2kV and
6kV settings), PPR is better. When the current is chopped
at an arbitrary time point (10kV setting), at the rising and
trailing edges, and at zero, PPR deteriorates. It still remains below 0.5% at maximum voltage and PRR, owing
to specifics of the used converter topology and high conversion frequency.
SEC.3
page 62
For 3 rail voltage settings, namely 460VDC, 520VDC and

590VDC, PPR was calculated by the formula, in Excel
convention,
,
where columns A-C contain each Vc values for 79 consecutive pulses, for 460, 520 and 590VDC, respectively. Alternatively, we varied the line voltage continuously from the
low to high level, looking for the least stable operation, i.e.,
for the largest Vc variation. For this method, PPR was calculated by (1) using VCmax, VCmin values from the whole
measurement range.
Short- and long-term PPR are plotted in Figure 9, Figure
10, respectively. The experimental curves shown in Figure
10 are calculated by (1), (2); they are marked as 3 rail experimental and overall experimental cont rails, respectively. The variability is larger than predicted by theory
accounting for the Factor 1 only (analytical curvesee
accompanying paper). This discrepancy can be attributed
to the measurement errors and propagation delays (Factors 2, 3).
Figure 8. Overlay of 80 frames (Vc - 100 V/div, I1 100 A/div) for:

a) high line, 2kV@1kHz; b) nominal line, 6kV@1kHz;
c) high line, 10kV@1kHz
Efficiency and Power Factor

The efficiency is calculated from the values of the input
and load power, the former being measured by a Voltech
PM300 power meter. Measuring the load power is indirect.
It is actually calculated as the energy per shot delivered to
the storage capacitor (E=Cs Vc2/2) multiplied by PRR. At
full power, the efficiency was about 92%, and power factor,
PF, was 94% (Figure 11). The efficiency values are lower
by 1-2% than expected and what could be deducted from
the loss estimation, and intuitively from the amount of the
dissipated heat. We note that the IGBTs baseplate overheat was less than 40C at all operational modes. One of
the possible sources of error is a low-accuracy Vc measurement (the probe P6015A is specified at 3% DC attenuation, excluding the oscilloscope error). Every percent of
voltage measurement error is translated to 2% of the energy measurement error, so the uncertainty of the efficiency measurement is quite pronounced.
With much smaller Cs=33nF the charge to 10kV is accomplished in 53s at low line, which allows PRR of 10kHz
with short-term- and long-term repeatability of 1.5% and
4.6%, respectively. However, the existing DSw limits the
operation to 1kHz CW.
62
T E C H N I C A L
R E S O U R C E S
ARTICLES

SEC.3
page 63
CONCLUSION
This development was a test case for low-cost general

technology of high repetition rate, high voltage, high
power, highly efficient capacitor charging. A crossover of
10kV, 20kJ/s, 1kHz PRR specifications was chosen for the
demonstration. An energy-dosing converter topology with
smart controls optimizing the switching frequency for high
efficiency was used. The switching is accomplished at a
frequency of up to 55kHz employing relatively slow inexpensive IGBTs. High efficiency allowed a compact all-air
cooled design. Excellent pulse-to-pulse repeatability was
demonstrated. As usual, the unit is protected against
short-circuit, arc, overvoltage, etc.
Fig. 9. Short-term repeatability
ACKNOWLEDGEMENT
The authors thank Mr. A. Lipovich for his help in

mechanical design.
REFERENCES
1.) B.D. Bedford and R.G. Hoft, Principles of Inverter Circuits,

Wiley, NY, 1964.
Fig. 10. Long-term repeatability as a function of charge voltage

summary of PSpice and analytical calculations and
experimental results.
2.) B. Kurchik, A. Pokryvailo and A. Schwarz, HV Converter for

Capacitor Charging, Pribory i Tekhnika Experimenta,
No. 4, pp. 121-124, 1990, Translation to English Plenum
Publishing Corp.
3.) M. Wolf and A. Pokryvailo, High Voltage Resonant Modular

Capacitor Charger Systems with Energy Dosage, Proc. 15th
IEEE Int. Conf. on Pulsed Power, Monterey CA, 13-17 June, 2005,
pp. 1029-1032.
4.) http://www.spellmanhv.com/Products/Rack-Supplies/SR.aspx
Fig. 11. Efficiency and power factor dependence on rail

voltage for several charge voltages.
63
T E C H N I C A L
R E S O U R C E S
GLOSSARY
page 64
AMPLIFIER, INVERTING
An amplifier whose output is 180 out of phase with its
input. Such an amplifier can be used with degenerative
feedback for stabilization purposes
ABSOLUTE ACCURACY
The correctness of the indicated value in terms of its deviation from the true or absolute value.
AMPLIFIER, NONINVERTING
An amplifier whose output is in phase with its input.
AC
In text, use lower case: ac. Abbreviation for
Alternating Current.
AMPLIFIER, OPERATIONAL
A dc amplifier whose gain is sufficiently large that its characteristics and behavior are substantially determined by its
input and feedback elements. Operational amplifiers are
widely used for signal processing and computational work.
AC BROWNOUT
The condition that exists when the ac line voltage drops
below some specified value.
ANODE
1) (electron tube or valve) An electrode through which a
principal stream of electrons leaves the interelectrode
space. 2) (semiconductor rectifier diode) The electrode
from which the forward current flows within the cell. (IEEE
Std 100-1988)
AC LINE
The set of conductors that route ac voltage from one point
to another.
AC LINE FILTER
A circuit filter placed in the ac line to condition or smooth
out variations that are higher in frequency than the line
frequency.
ANSI
Abbreviation for American National Standards Institute
ALTERNATING CURRENT
(ac) A periodic current the average value of which over a
period is zero. Unless distinctly specified otherwise, the
term refers to a current which reverses at regularly recurring intervals of time and which has alternately positive
and negative values.
APPARENT POWER
Power value obtained in an ac circuit as the product of
current times voltage.
ARC
A discharge of electricity through a gas, normally characterized by a voltage drop in the immediate vicinity of the
cathode approximately equal to the ionization potential of
the gas. (IEE Std 100-1988)
AMBIENT TEMPERATURE
The average temperature of the environment immediately
surrounding the power supply. For forced air-cooled units,
the ambient temperature is measured at the air intake.
See also Operating Temperature, Storage Temperature,
Temperature Coefficient.
ASYMMETRICAL WAVEFORM
A current or voltage waveform that has unequal excursions above and below the horizontal axis.
AMPERE
(A) Electron or current flow representing the flow of one
coulomb per second past a given point in a circuit.
ATTENUATION
Decrease in amplitude or intensity of a signal.
AUTHORIZED PERSON
A qualified person who, by nature of his duties or occupation, is obliged to approach or handle electrical equipment
or, a person who, having been warned of the hazards involved, has been instructed or authorized to do so by
someone in authority.
AMPLIFIER
A circuit or element that provides gain.
AMPLIFIER, DC
A direct coupled amplifier that can provide gain for zerofrequency signals.
AMPLIFIER, DIFFERENTIAL
An amplifier which has available both an inverting and
a noninverting input, and which amplifies the difference
between the two inputs.
SEC.4
64
T E C H N I C A L
R E S O U R C E S
GLOSSARY
AUTO TRANSFORMER
A single winding transformer with one or more taps.
page 65
BIAS SUPPLY
Power source fitted with output controls, meters,
terminals and displays for experimental bench top
use in a laboratory.
AUTOMATIC CROSSOVER
The characteristic of a power supply having the capability
of switching its operating mode automatically as a function
of load or setting from the stabilization of voltage to the
stabilization of current. The term automatic crossover
power supply is reserved for those units having substantially equal stabilization for both voltage and current. Not
used for voltage-limited current stabilizers or current-limited voltage stabilizers. See also CROSSOVER POINT.
BIFILAR WINDING
Two conductors wound in parallel.
BIPOLAR
Having two poles, polarities or directions.
BIPOLAR PLATE
An electrode construction where positive and negative
active materials are on opposite sides of an electronically
conductive plate.
AUTOMATIC GAIN CONTROL (AGC)

A process or means by which gain is automatically
adjusted in a specified manner as a function of input or
other specified parameters. (IEEE Std 100-1988)
BIPOLAR POWER SUPPLY

A special power supply which responds to the sense as
well as the magnitude of a control instruction and is able
to linearly pass through zero to produce outputs of either
positive or negative polarity.
AUXILIARY SUPPLY
A power source supplying power other than load power as
required for the proper functioning of a device.
BIT
A binary unit of digital information having a value of "0" or
"1". See also Byte.
AWG
Abbreviation for American Wire Gauge.
BLACK BOX
Element in a system specified by its function, or operating
characteristics.
BANDWIDTH
Based on the assumption that a power supply can be
modeled as an amplifier, the bandwidth is that frequency
at which the voltage gain has fallen off by 3 dB.
Bandwidth is an important determinant of transient
response and output impedance.
BLEED
A low current drain from a power source.
BLEED RESISTOR
A resistor that allows a small current drain on a
power source to discharge filter capacitors or to
stabilize an output.
BASEPLATE TEMPERATURE
The temperature at the hottest spot on the mounting
platform of the supply.
BOBBIN
1) A non-conductive material used to support windings.
2) A cylindrical electrode (usually the positive) pressed
from a mixture of the active material, a conductive material, such as carbon black, the electrolyte and/or binder
with a centrally located conductive rod or other means
for a current collector.
BEAD
A small ferrite normally used as a high frequency
inductor core.
BEAM SUPPLY
Power supply which provides the accelerating energy for
the electrons or ions.
BENCH POWER SUPPLY

Power source fitted with output controls, meters,
terminals and displays for experimental bench top
use in a laboratory.
SEC.4
BODE PLOT
A plot of gain versus frequency for a control loop. It usually
has a second plot of phase versus frequency.
65
BOOST REGULATOR
One of several basic families of switching power supply
topologies. Energy is stored in an inductor during the
pulse then released after the pulse.
T E C H N I C A L
SEC.4
R E S O U R C E S
GLOSSARY
BREAKDOWN VOLTAGE
1)The voltage level which causes insulation failure.
2) The reverse voltage at which a semiconductor device
changes its conductance characteristics.
CAPACITANCE
Inherent property of an electric circuit or device that
opposes change in voltage. Property of circuit whereby
energy may be stored in an electrostatic field.
BRIDGE CIRCUIT
Circuit with series parallel groups of components.
CAPACITANCE-DISTRIBUTED
The capacitance in a circuit resulting from adjacent turns
on coils, parallel leads and connections.
BRIDGE CONVERTER
A power conversion circuit with the active elements
connected in a bridge configuration.
CAPACITIVE COUPLING
Coupling resulting from the capacitive effect
between circuit elements.
BRIDGE RECTIFIER
Full-wave rectifier circuit employing two or more rectifiers
in a bridge configuration.
CAPACITANCE, DISTRIBUTED
The current flow between segregated conductive metal
parts; voltage and frequency dependent.
BROWNOUT
The condition created during peak usage periods when
electric utility companies intentionally reduce their line
voltage by approximately 10 to 15 percent to counter
excessive demand.
CAPACITOR
A device that stores a charge. A simple capacitor consists
of two conductors separated by a dielectric.A device that
stores a charge. A simple capacitor consists of two conductors separated by a dielectric.
BUCK REGULATOR
The condition created during peak usage periods when
electric utility companies intentionally reduce their line
voltage by approximately 10 to 15 percent to counter
excessive demand.
CAPACITOR INPUT FILTER

Filter employing capacitor as its input.
BUFFER
An isolating circuit used to prevent a driven circuit from
influencing a driving circuit. (IEEE Std 100-1988)
CATHODE
1) (electron tube or valve) An electrode through which a
primary stream of electrons enters the interelectrode
space. 2) (semiconductor rectifier diode) The electrode to
which the forward current flows within the cell. (IEEE Std
100-1988).
BUFFER
The energy storage capacitor at the front end
of a regulator.
CATHODE RAY TUBE (CRT)

A display device in which controlled electron beams are
used to present alphanumeric or graphical data on an
electroluminescent screen. (IEEE Std 100-1988).
BULK VOLTAGE
The energy storage capacitor at the front
end of a regulator.
BURN IN
The operation of a newly fabricated device or system prior
to application with the intent to stabilize the device, detect
defects, and expose infant mortality.
CATHODE RAY TUBE

An electron-beam tube in which the beam can be focused
to a small cross section on a luminescent screen and varied in position and intensity to produce a visible pattern.
(IEEE Std 100-1988).
BUS
The common primary conductor of power from a power
source to two or more separate circuits.
BYTE
A sequence of binary digits, frequently comprised of eight
(8) bits, addressed as a unit. Also see BIT.
page 66
CENTER TAP
Connection made to center of an electronic device.
66
CGS UNIT
Abbreviation for the Centimeter-Gram Second
Unit of measurement.
T E C H N I C A L
R E S O U R C E S
GLOSSARY
CHARGE
1) The conversion of electrical energy, provided in the
form of a current from an external source, into chemical
energy within a cell or battery. 2) The potential energy
stored in a capacitive electrical device.
SEC.4
page 67
COMMON CHOKE
See INTEGRATED MAGNETICS.
COMMON-MODE NOISE
The component of noise voltage that appears equally and
in phase on conductors relative to a common reference.
CHASSIS
The structure supporting or enclosing the power supply.
COMMON-MODE OUTPUT
That electrical output supplied to an impedance connected
between the terminals of the ungrounded floating output of
a power supply, amplifier, or line-operated device, and the
ground point to which the source power is returned.
CHASSIS GROUND
The voltage potential of the chassis.
CHOKE COIL
An inductor.
COMMON POINT
With respect to operationally programmable power supplies one output/sense terminal is designated "common"
to which load, reference and external programming signal
all return.
CHOKE, RF
A choke coil with a high impedance at radio frequencies.
CIRCUIT INPUT FILTER

A filter employing an inductor (L) or an inductor/capacitor
(L/C) as its input.
COMMON RETURN
A return conductor common to two or more circuits.
CIRCULAR MIL
Cross-sectional area of a conductor one mil in diameter.
COMPARISON AMPLIFIER
A dc amplifier which compares one signal to a stable
reference, and amplifies the difference to regulate the
power supply power-control elements.
CIRCULATING CURRENT
See GROUND LOOP.
CLAMP DIODE
A diode in either a clipper or clamp circuit.
COMPENSATION
The addition of circuit elements to assist in stabilization of
a control loop.
CLOSED LOOP CONTROL

A type of automatic control in which control actions are
based on signals fed back from the controlled equipment
or system. (IEEE Std 100-1988)
COMPLIANCE
Agency certification that a product meets its standards.
See also SAFETY COMPLIANCE.
CLIPPER CIRCUIT
A circuit that blocks or removes the portion of a voltage
waveform above some threshold voltage.
COMPLIMENTARY TRACKING
A system of interconnection of two voltage stabilizers
by which one voltage (the slave) tracks the other
(the master).
CLOSED-LOOP CONTROL SYSTEM

(control system feedback) A control system in which the
controlled quantity is measured and compared with a
standard representing the desired performance.
Note: Any deviation from the standard is fed back into
the control system in such a sense that it will reduce the
deviation of the controlled quantity from the standard.
(IEEE Std 100-1988)
COLLECTOR
1) Electronic connection between the electrochemical cell
electrode and the external circuit. 2) In a transistor, the
semiconductor section which collects the majority carriers.
COMPLIANCE VOLTAGE
The output dc voltage of a constant current supply.
COMPLIANCE RANGE
Range of voltage needed to sustain a given constant
current throughout a range of load resistance.
67
CONDUCTANCE (G)
The ability to conduct current. It is equal to amperes per
volt, or the reciprocal of resistance, and is measured in
siemens (metric) or mhos (English). G = 1/R.
T E C H N I C A L
R E S O U R C E S
GLOSSARY
CONSTANT CURRENT LIMITING CIRCUIT

Current-limiting circuit that holds output current at some
maximum value whenever an overload of any magnitude
is experienced.
page 68
CORONA
1) (air) A luminous discharge due to ionization of the air
surrounding a conductor caused by a voltage gradient exceeding a certain critical value. 2) (gas) A discharge with
slight luminosity produced in the neighborhood of a conductor, without greatly heating it, and limited to the region
surrounding the conductor in which the electric field exceeds a certain value. 3) (partial discharge) (corona measurement) A type of localized discharge resulting from
transient gaseous ionization in an insulation system when
the voltage stress exceeds a critical value. The ionization
is usually localized over a portion of the distance between
the electrodes of the system. (IEEE Std 100-1988)
CONSTANT VOLTAGE CHARGE

A charge during which the voltage across the battery
terminals is maintained at a steady state.
CONTINUOUS DUTY
A requirement of service that demands operation at a
substantially constant load for an indefinitely long time.
See also INTERMITTENT DUTY.
CONTROL GRID
A grid, ordinarily placed between the cathode
and an anode, for use as a control electrode.
(IEEE Std 100-1988)
CORONA EXTINCTION VOLTAGE

(CEV) (corona measurement) The highest voltage at
which continuous corona of specified pulse amplitude no
longer occurs as the applied voltage is gradually decreased from above the corona inception value. Where
the applied voltage is sinusoidal, the CEV is expressed as
0.707 of the peak voltage. (IEEE Std 100-1988)
CONTROL LOOP
A feedback circuit used to control an output signal.
See also LOOP.
CORONA INCEPTION VOLTAGE

(CIV) (corona measurement) The lowest voltage at which
continuous corona of specified pulse amplitude occurs as
the applied voltage is gradually increased. Where the applied voltage is sinusoidal, the CIV is expressed as 0.707
of the peak voltage. (IEEE Std 100-1988)
CONTROL RANGE
The parameter over which the controlled signal maybe
adjusted and still meet the unit specifications.
CONTROL REMOTE
Control over the stabilized output signal by means located
outside or away from the power supply. May or may not be
calibrated.
CREEPAGE
The movement of electrolyte onto surfaces of electrodes
or other components of a cell with which it is not normally
in contact.
CONTROL RESOLUTION
The smallest increment of the stabilized output signal that
can be reliably repeated.
CREEPAGE DISTANCE
The shortest distance separating two conductors as
measured along a surface touching both conductors.
CONVECTION-COOLED POWER SUPPLY

A power supply cooled exclusively from the natural
motion of a gas or a liquid over the surfaces of heat
dissipating elements.
CROSS-REGULATION
In a multiple output power supply, the percent voltage
change at one output caused by the load change on
another output.
CONVERTER
A device that changes the value of a signal or quantity.
Examples: DC-DC; a device that delivers dc power when
energized from a dc source. Fly-Back; a type of switching
power supply circuit. See also FLYBACK CONVERTER.
Forward; a type of switching power supply circuit.
See also FORWARD CONVERTER.
CORE
Magnetic material serving as a path for magnetic flux.
SEC.4
68
CROSSOVER POINT
That point on the operating locus of a voltage/current automatic crossover power supply formed by the intersection
of the voltage-stabilized and current-stabilized output
lines. The resistance value (E/I) defined by this intersection is the matching impedance of the power supply, which
will draw the maximum output power.
See also AUTOMATIC CROSSOVER.
T E C H N I C A L
R E S O U R C E S
GLOSSARY
CROSSOVER, VOLTAGE/CURRENT
Voltage/Current crossover is that characteristic of a power
supply that automatically converts the mode of operation
from voltage regulation to current regulation (or vice
versa) as required by preset limits.
page 69
DC-DC CONVERTER
A circuit or device that changes a dc input signal value to
a different dc output signal value.
DECAY TIME
See FALL TIME
CROWBAR
An overvoltage protection circuit which rapidly places a
low resistance shunt across the power supply output terminals if a predetermined voltage is exceeded.
DERATING
(reliability) The intentional reduction of stress/strength
ratio in the application of an item, usually for the purpose
of reducing the occurrence of stress-related failures.
(IEEE Std 100-1988)
CSA
Abbreviation for Canadian Standards Association.
DIELECTRIC
An insulating material between conductors.
CURRENT CONTROL
See CURRENT STABILIZATION
DIELECTRIC CONSTANT (K)

For a given dielectric material, the ratio of the value of a
capacitor using that material to the value of an equivalent
capacitor using a standard dielectric such as dry air or a
vacuum.
CURRENT FOLDBACK
See FOLDBACK CURRENT LIMITING.
CURRENT LIMIT KNEE

The point on the plot of current vs voltage of a supply at
which current starts to foldback, or limit.
DIELECTRIC WITHSTAND VOLTAGE

Voltage an insulating material will withstand before
flashover or puncture.
See also HI-POT TEST, ISOLATION.
CURRENT LIMITING
An electronic overload protection circuit which limits the
maximum output current to a preset value.
DIFFERENTIAL VOLTAGE
The difference in voltages at two points as measured with
respect to a common reference.
CURRENT MODE
The functioning of a power supply so as to produce
a stabilized output current.
DRIFT
A change in output over a period of time independent of
input, environment or load
CURRENT SENSING RESISTOR

A resistor placed in series with the load to develop a
voltage proportional to load current.
DRIVER
A current amplifier used for control of another
device or circuit.
CURRENT SOURCE
A power source that tends to deliver constant current.
DUTY CYCLE
1) The ratio of time on to time off in a recurring event. 2)
The operating regime of a cell or battery including factors
such as charge and discharge rates, depth of discharge,
cycle length and length of time in the standby mode.
CURRENT STABILIZATION
The process of controlling an output current.
D
DYNAMIC FOCUS
A means of modulating the focus voltage as a function of
the beam position. (Bertan High Voltage)
DC
In text, use lower case: dc. Abbreviation for Direct Current.
DC COMPONENT
The dc value of an ac wave that has an axis
other than zero.
SEC.4
69
DYNAMIC LOAD
A load that rapidly changes from one level to another. To
be properly specified, both the total change and the rate of
change must be stated.
T E C H N I C A L
R E S O U R C E S
GLOSSARY
page 70
ELECTRONIC
Of, or pertaining to, devices, circuits, or systems utilizing
electron devices. Examples: Electronic control, electronic
equipment, electronic instrument, and electronic circuit.
(IEEE Std 100-1988)
EARTH
An electrical connection to the earth frequently using a
grid or rod(s). See also GROUND.
E-BEAM
Electron Beam. (Bertan High Voltage)
ELECTRONIC LOAD
A test instrument designed to draw various and specified
amounts of current or power from a power source.
EDDY CURRENTS
A circulating current induced in a conducting material by a
varying magnetic field.
ELECTRON VOLT
A measure of energy. The energy acquired by an electron
passing through a potential of one volt.
EFFECTIVE VALUE
The value of a waveform that has the equivalent heating
effect of a direct current. For sine waves, the value is .707
X Peak Value; for non-sinusoidal waveforms, the Effective
Value = RMS (Root Mean Square) Value.
ELECTROPHORESIS
A movement of colloidal ions as a result of the application
of an electric potential. (IEEE Std 100-1988)
EMF
Abbreviation for Electromotive Force.
EFFICIENCY
1) The ratio of total output power to total input power,
expressed as a percentage, under specified conditions.
2) The ratio of the output of a secondary cell or battery on
discharge to the input required to restore it to the initial
state of charge under specified conditions.
EMI
Abbreviation for Electromagnetic Interference.
EMI FILTER
A circuit composed of reactive and resistive components
for the attenuation of radio frequency components being
emitted from a power supply. See also EMI.
ELECTRIC
Containing, producing, arising from, actuated by, or carrying electricity, or designed to carry electricity and capable
of so doing. Examples: Electric eel, energy, motor, vehicle,
wave. Note: Some dictionaries indicate electric and electrical as synonymous, but usage in the electrical engineering
field has in general been restricted to the meaning given in
the definitions above. It is recognized that there are borderline cases wherein the usage determines the selection.
See ELECTRICAL. (IEEE Std 100-1988)
EMI FILTERING
Process or network of circuit elements to reduce electromagnetic interference emitted from or received by an electronic device. See also EMI.
EMISSION
1) (laser-maser) The transfer energy from matter to a radiation field. 2) (radio-noise emission) An act of throwing out
or giving off, generally used here in reference to electromagnetic energy. (IEEE Std 100-1988)
ELECTRICAL
(general) Related to, pertaining to, or associated with
electricity but not having its properties or characteristics.
Examples: Electrical engineer, handbook, insulator, rating,
school, unit.
EMISSION CURRENT
The current resulting from electron emission.
(IEEE Std 100-1988)
ELECTRON BEAM
A collection of electrons which may be parallel,
convergent, or divergent. (Bertan High Voltage)
EQUIVALENT CIRCUIT
An electrical circuit that models the fundamental properties
of a device or circuit.
ELECTRON (e-)
Negatively charged particle.
ELECTRON GUN
(electron tube) An electrode structure that produces and
may control, focus, deflect, and converge one or more
electron beams. (IEEE Std 100-1988)
SEC.4
70
EQUIVALENT LOAD
An electrical circuit that models the fundamental
properties of a load.
T E C H N I C A L
R E S O U R C E S
GLOSSARY
EQUIVALENT SERIES INDUCTANCE (ESI)

The amount of inductance in series with an ideal
capacitor which exactly duplicates the performance
of a real capacitor.
page 71
FEEDBACK
The process of returning part of the output signal of a system to its input.
FEED FORWARD
A control technique whereby the line regulation of a power
supply is improved by directly sensing the input voltage.
EQUIVALENT SERIES RESISTANCE (ESR)

The amount of resistance in series with an ideal
capacitor which exactly duplicates the performance
of a real capacitor.
FEED THROUGH
A plated-through hole in a printed circuit board which electrically connects a trace on top of the board with a trace on
the bottom side.
ERROR AMPLIFIER
An operational amplifier, or differential amplifier, in a control loop that produces an error signal whenever a sensed
output differs from a reference voltage.
FERRITE
A ceramic material that exhibits low loss at high frequencies, and which contains iron oxide mixed with oxides or
carbonates of one or more metals such as manganese,
zinc, nickel or magnesium.
ERROR SIGNAL
The output voltage of an error amplifier produced by the
difference between the reference and the input signal
times the gain of the amplifier.
FET
Abbreviation for Field Effect Transistor.
ERROR VOLTAGE
The output voltage of the error amplifier in a control loop.
FIELD EFFECT TRANSISTOR (FET)

Transistor in which the resistance of the current path from
source to drain is modulated by applying a transverse
electric field between two electrodes. See also JUNCTIONFIELD EFFECT TRANSISTOR, METAL OXIDE,
SEMICONDUCTOR FIELD EFFECT TRANSISTOR.
ESD
Abbreviation for Electrostatic Discharge.
ESL
Abbreviation for Equivalent Series Inductance.
ESR
Abbreviation for Equivalent Series Resistance.
FIELD EMISSION
Electron emission from a surface due directly to high voltage gradients at the emitting surface. (IEEE Std 100-1988)
FIELD EMISSION GUN

An electron gun with an extractor electrode which pulls or
extracts electrons off the filament.
FAILURE MODE
The way in which a device has ceased to meet specified
minimum requirements.
FILAMENT
(electron tube) A hot cathode, usually in the form of a wire
or ribbon, to which heat may be supplied by passing current through it. Note: This is also known as a filamentary
cathode. (IEEE Std 100-1988)
FALL TIME
The time required for a pulse to decrease from
90 percent to 10 percent of its maximum positive
(negative) amplitude.
FILAMENT CURRENT
The current supplied to a filament to heat it.
(IEEE Std 100-1984)
FAN COOLED
A method of forced-air cooling used to maintain design.
FARAD
Unit of measurement of capacitance. A capacitor has a
capacitance of one farad when a charge of one coulomb
raises its potential one volt: C = Q/E.
SEC.4
71
FILAMENT OUTPUT
Power supply which heats the filament of an electron
column, CRT or x-ray tube. In some applications,
the filament output "floats" on the accelerating voltage.
(Bertan High Voltage)
T E C H N I C A L
SEC.4
R E S O U R C E S
GLOSSARY
FILAMENT VOLTAGE
Power supply which heats the filament of an electron
column, CRT or x-ray tube. In some applications,
the filament output "floats" on the accelerating voltage.
page 72
FOLDBACK CURRENT LIMITING

A power supply output protection circuit whereby the output current decreases with increasing overload, reaching a
minimum at short circuit. This minimizes the internal power
dissipation under overload conditions. Foldback current
limiting is normally used with linear regulators
FILTER
One or more discrete components positioned in a circuit to
attenuate signal energy in a specified band of frequencies.
FORWARD CONVERTER
A power supply switching circuit that transfers energy
to the transformer secondary when the switching transistor
is on.
FLASHOVER
1) (general) A disruptive discharge through air around or
over the surface of solid or liquid insulation, between parts
of different potential or polarity, produced by the application of voltage wherein the breakdown path becomes sufficiently ionized to maintain an electric arc. 2) (high voltage
ac cable termination) A disruptive discharge around or
over the surface of an insulating member, between parts
of different potential or polarity, produced by the application of voltage wherein the breakdown path becomes sufficiently ionized to maintain an electric arc. 3) (high voltage
testing) Term used when a disruptive discharge occurs
over the surface of a solid dielectric in a gaseous or liquid
medium. (IEEE Std 100-1988)
FREE WHEEL DIODE

A diode in a pulse-width modulated switching power
supply that provides a conduction path for the counter
electromotive force of an output choke.
FREQUENCY
Number of cycles per second (measured in Hertz).
FULL BRIDGE CONVERTER

A power switching circuit in which four power switching
devices are connected in a bridge configuration to drive a
transformer primary.
FLOATING NETWORK OR COMPONENTS

A network or component having no terminal at ground
potential. (IEEE Std 100-1988)
FULL BRIDGE RECTIFIER

A rectifier circuit that employs four diodes per phase.
FULL WAVE RECTIFIER

Rectifier circuit that produces a dc output for each half
cycle of applied alternating current.
FLOATING OUTPUT
Ungrounded output of a power supply where either output
terminal may be referenced to another specified voltage.
FUSE
Safety protective device that permanently opens an
electric circuit when overloaded. See also OVERCURRENT DEVICE, OVERCURRENT PROTECTIVE DEVICE.
FLYBACK CONVERTER
A power supply switching circuit which normally uses
a single transistor. During the first half of the switching
cycle the transistor is on and energy is stored in a
transformer primary; during the second half of the
switching cycle this energy is transferred to the
transformer secondary and the load.
GAIN
Ratio of an output signal to an input signal. See also
CLOSED LOOP GAIN, GAIN MARGIN, OPEN LOOP
GAIN.
FOCUS
(oscillograph) Maximum convergence of the electron
beam manifested by minimum spot size on the phosphor
screen. (IEEE Std 100-1988)
GAUSS
Measure of flux density in Maxwells per square centimeter
of cross-sectional area. One Gauss is 10-4 Tesla
FOCUSING ELECTRODE
(beam tube) An electrode the potential of which is adjusted
to focus an electron beam. (IEEE Std 100-1988)
73
GLITCH
1) An undesired transient voltage spike occurring
on a signal. 2) A minor technical problem arising in
electrical equipment.
T E C H N I C A L
SEC.4
R E S O U R C E S
GLOSSARY
GPIB
General purpose interface bus, also known as IEEE-488.
HENRY (H)
Unit of measurement of inductance. A coil has one henry
of inductance if an EMF of one volt is induced when current through an inductor is changing at rate of one ampere
per second
GRID
1) In batteries, a framework for a plate or electrode which
supports or retains the active materials and acts as a current collector. 2) In vacuum tubes, an element used to
control the flow of electrons. 3) A network of equally
spaced parallel lines, one set spaced perpendicular to
the other.
HERTZ (Hz)
The SI unit of measurement for frequency, named in honor
of Heinrich Hertz who discovered radio waves. One hertz
equals one cycle per second.
HICCUP
A transient condition that momentarily confuses
a control loop.
GROUND
A conducting connection, whether intentional or accidental, by which an electric circuit or equipment is connected
to earth, or to some conducting body that serves in place
of earth. (National Electric Code)
HIGH LINE
Highest specified input operating voltage.
GROUND BUS
A bus to which individual grounds in a system are attached
and that in turn is grounded at one or more points.
HIGH VOLTAGE ASSEMBLY

The portion of a high voltage power supply which contains
the high voltage circuits which are critical to the performance and reliability of a high voltage power supply.
GROUNDED
Connected to or in contact with earth or connected to
some extended conductive body which serves instead of
the earth.
HI-POT TEST (HIGH POTENTIAL TEST)

A test performed by applying a high voltage for a specified
time to two isolated points in a device to determine adequacy of insulating materials.
GROUND LOOP
A condition that causes undesirable voltage levels when
two or more circuits share a common electrical return or
ground lines.
HOLDING TIME
See HOLDUP TIME
HOLDUP TIME
The time under worst case conditions during which a
power supply's output voltage remains within specified limits following the loss or removal of input power. Sometimes
called Holding Time or Ride-Through.
HALF-BRIDGE CONVERTER
A switching power supply design in which two power
switching devices are used to drive the transformer
primary. See also BRIDGE RECTIFIER.
HYBRID SUPPLIES
A power supply that combines two or more different regulation techniques, such as ferroresonant and linear or
switching and linear, or one that takes advantage of hybrid
technology.
HALF-WAVE RECTIFIER
A circuit element, such as a diode, that rectifies only onehalf the input ac wave to produce a pulsating dc output.
HEADROOM
The difference between the bulk voltage and the output
voltage in a linear series pass regulator.
See also DIFFERENTIAL VOLTAGE.
HEAT SINK
The medium through which thermal energy is dissipated.
page 74
I-BEAM
Ion Beam. (Bertan High Voltage)
74
IC
Abbreviation for Integrated Circuit.
T E C H N I C A L
SEC.4
R E S O U R C E S
GLOSSARY
page 75
IEC
Abbreviation for International Electrotechnical
Commission.
INVERTER
1) A device that changes dc power to ac power. 2) A circuit, circuit element or device that inverts the input signal.
IMPEDANCE (Z)
Total resistance to flow of an alternating current as a result
of resistance and reactance.
ION GUN
A device similar to an electron gun but in which the
charged particles are ions. Example: proton gun.
(IEEE Std 100-1988)
IEEE
Abbreviation for Institute of Electrical and Electronics
Engineers.
ION BEAM
A collection of ions which may be parallel, convergent, or
divergent. (Bertan High Voltage)
INDUCED CURRENT
Current that flows as a result of an Induced EMF
(Electromotive Force).
ISOLATION
The electrical separation between two circuits, or
circuit elements.
INDUCED EMF
Voltage induced in a conductor in a varying magnetic field.
ISOLATION TRANSFORMER
A transformer with a one-to-one turns ratio. See also
STEP-DOWN TRANSFORMER STEP-UP TRANSFORMER, TRANSFORMER
INPUT
The ability to turn off the output of a power supply from a
remote location
ISOLATION VOLTAGE
The maximum ac or dc specified voltage that may be
continuously applied between isolated circuits.
INDUCED IMPEDANCE
The impedance of the input terminals of a circuit or device,
with the input disconnected.
INDUCED FILTER
A low-pass or band-reject filter at the input of a power
supply which reduces line noise fed to the supply. This
filter may be external to the power supply.
JOULE (J)
Unit of energy equal to one watt-second.
INDUCED SURGE
See INRUSH CURRENT
INPUT VOLTAGE RANGE

The range of input voltage values for which a power
supply or device operates within specified limits.
KELVIN (K)
1) Unit of temperature in the International System of Units
(Sl) equal to the fraction 1/273.16 of the thermodynamic
temperature of the triple point of water. The kelvin temperature scale uses Celsius degrees with the scale shifted by
273.16. Therefore, 0 K is at absolute zero. Add 273.16 to
any Celsius value to obtain the corresponding value in
kelvins. 2) A technique using 4 terminals to isolate current
carrying leads from voltage measuring leads.
INRUSH CURRENT
The range of input voltage values for which a power
supply or device operates within specified limits.
INSTANTANEOUS VALUE
The measured value of a signal at a given moment in time.
INSULATION
Non-conductive materials used to separate
electric circuits.
INSULATION RESISTANCE
The resistance offered, usually measured in megohms, by
an insulating material to the flow of current resulting from
an impressed dc voltage
KIRCHOFF'S CURRENT LAW

At any junction of conductors in a circuit, the algebraic
sum of the current is zero
75
KIRCHOFF'S VOLTAGE LAW

In a circuit, the algebraic sum of voltages around the
circuit is equal to zero.
T E C H N I C A L
R E S O U R C E S
GLOSSARY
page 76
LINEAR SUPPLY REGULATION

An electronic power supply employing linear regulation
techniques. See also LINEAR REGULATION.
LATCH-UP
A part of the control circuit for a power supply that goes
into a latched condition.
LINE CONDITIONER
A circuit or device designed to improve the quality
of an ac line.
L-C FILTER
A low pass filter that consists of an inductance (L) and a
capacitance (C). Also known as an averaging filter.
LINE EFFECT
See LINE REGULATION.
LEAKAGE CURRENT
1) The ac or dc current flowing from input to output and/or
chassis of an isolated device at a specified voltage.
2) The reverse current in semiconductor junctions.
LINE REGULATION
A regulation technique wherein the control device, such as
transistor, is placed in series or parallel with the load. Output is regulated by varying the effective resistance of the
control device to dissipate unused power.
See also LINEAR SUPPLY, REGULATION.
LED
Symbol for Light-Emitting Diode.
LINE REGULATOR
Power conversion equipment that regulates and/or
changes the voltage of incoming power.
LINE
1) Medium for transmission of electricity between circuits
or devices. 2) The voltage across a power transmission
line. See also HIGH LINE, LOW LINE.
LINE TRANSIENT
A perturbation outside the specified operating range of an
input or supply voltage.
LINEAR
1) In a straight line. 2) A mathematical relationship in
which quantities vary in direct proportion to one another,
the result of which, when plotted, forms a straight line.
LOAD
Capacitance, resistance, inductance or any combination
thereof, which, when connected across a circuit determines current flow and power used.
LINEARITY
1) The ideal property wherein the change in the value of
one quantity is directly proportional to the change in the
value of another quantity, the result of which, when plotted
on graph, forms a straight line. 2) Commonly used in reference to Linearity Error.
LOAD DECOUPLING
The practice of placing filter components at the load
to attenuate noise.
LOAD EFFECTS
See LOAD REGULATION
LINEAR SUPPLY REGULATION

The deviation of the output quantity from a
specified reference line.
LOAD IMPEDANCE
The complex resistance to the flow of current posed
by a load that exhibits both the reactive and resistive
characteristics.
LINEAR PASS
See SERIES PASS
LINEAR REGULATION
A regulation technique wherein the control device, such as
transistor, is placed in series or parallel with the load. Output is regulated by varying the effective resistance of the
control device to dissipate unused power.
See also LINEAR SUPPLY, REGULATION.
LINEAR REGULATOR
A power transformer or a device connected in series with
the load of a constant voltage power supply in such a way
that the feedback to the series regulator changes its voltage drop as required to maintain a constant dc output.
SEC.4
76
LOAD REGULATION
1) Static: The change in output voltage as the load is
changed from specified minimum to maximum and maximum to minimum, with all other factors held constant. 2)
Dynamic: The change in output voltage expressed as a
percent for a given step change in load current. Initial and
final current values and the rates of change must be specified. The rate of change shall be expressed as current/unit
of time, e.g., 20 amperes A/ second. The dynamic regulation is expressed as a percent for a worst case peak-topeak deviation for dc supplies, and worst case rms
deviation for ac supplies.
T E C H N I C A L
SEC.4
R E S O U R C E S
GLOSSARY
LOCAL CONTROL
Control over the stabilized output signal by means
located within or on the power supply. May or may
not be calibrated.
MAINS
The utility AC power source.
LOCAL SENSING
Using the power supply output voltage terminals as the
error-sensing points to provide feedback to the voltage
regulator.
page 77
M
MASTER-SLAVE OPERATION
A method of interconnecting two or more supplies such
that one of them (the master) serves to control the others
(the slaves). The outputs of the slave supplies always remain equal to or proportional to the output of the master
LOGIC HIGH
A voltage representing a logic value of one
(1) in positive logic.
MAXIMUM LOAD
1) The highest allowable output rating specified for any or
all outputs of a power supply under specified conditions including duty cycle, period and amplitude. 2) The highest
specified output power rating of a supply specified under
worst case conditions.
LOGIC INHIBIT/ENABLE
A referenced or isolated logic signal that turns a power
supply output off or on.
LOGIC LOW
A voltage representing a logic value of zero
(0) in positive logic.
MINIMUM LOAD
1) The lowest specified current to be drawn on a constant
voltage power supply for the voltage to be in a specified
range. 2) For a constant current supply, the maximum
value of load resistance.
LONG-TERM STABILITY
The output voltage change of a power supply, in percent,
due to time only, with all other factors held constant. Longterm stability is a function of component aging.
MODULAR
1) A physically descriptive term used to describe a power
supply made up of a number of separate subsections,
such as an input module, power module, or filter module.
2) An individual power unit patterned on standard dimensions and capable of being integrated with other parts or
units into a more complex and higher power system.
LOOP
The path used to circulate a signal. See also CLOSED
LOOP, CONTROL LOOP, OPEN LOOP.
LOOP GAIN
The ratio of the values of a given signal from one point to
another in a loop. See also GAIN.
MODULATOR
The control element of a switching power supply.
LOOP RESPONSE
The speed with which a loop corrects for specified
changes in line or load.
MOSFET
Abbreviation for Metal Oxide Semiconductor Field Effect
Transistor.
LOOP STABILITY
A term referencing the stability of a loop as measured
against some criteria, e.g., phase margin and gain margin.
MTBF
Abbreviation for Mean Time Between Failure.
LOW LINE
Lowest specified input operating voltage.
77
NEGATIVE FEEDBACK:
1) (circuits and systems) The process by which part of the
signal in the output circuit of an amplifying device reacts
upon the input circuit in such a manner as to counteract
the initial power, thereby decreasing the amplification. 2)
(control) (industrial control) A feedback signal in a direction
to reduce the variable that the feedback represents. 3)
T E C H N I C A L
R E S O U R C E S
GLOSSARY
(degeneration) (stabilized feedback) (data transmission)

The process by which a part of the power in the output circuit of an amplifying device reacts upon the input circuit in
such a manner as to reduce the initial power, thereby reducing the amplification. (IEEE Std 100-1988)
page 78
OFFSET VOLTAGE
The dc voltage that remains between the input terminals of
a dc amplifier when the output current voltage is zero
OHM
Unit of measure of resistance
NEGATIVE RAIL
The more negative of the two conductors at the output of a
power supply.
OP-AMP
Abbreviation for Operational Amplifier
OHM
The difference in potential between the terminals of a cell
or voltage when the circuit is open (no-load condition).
See NO LOAD VOLTAGE.
NEGATIVE REGULATOR
A voltage regulator whose output voltage is negative compared to the voltage at the return.
NEGATIVE TEMPERATURE COEFFICIENT
A decreasing function with increasing temperature. The
function may be resistance, capacitance, voltage, etc.
OPEN-FRAME CONSTRUCTION
A construction technique where the supply is not provided
with an enclosure.
NODE
The junction of two or more branches in a circuit.
OPEN LOOP
A signal path without feedback.
NOISE
The aperiodic random component on the power source
output which is unrelated to source and switching frequency. Unless specified otherwise, noise is expressed in
peak-to-peak units over a specified bandwidth.
OPEN LOOP GAIN

Ratio of output signal to input signal without feedback.
OPERATING TEMPERATURE RANGE

The range of ambient, baseplate or case temperatures
through which a power supply is specified to operate
safely and to perform within specified limits. See also AMBIENT TEMPERATURE, STORAGE TEMPERATURE.
NO LOAD VOLTAGE
Terminal voltage of battery or supply when no current is
flowing in external circuit. See OPEN CIRCUIT VOLTAGE
OPERATIONAL AMPLIFIER (OP-AMP)

A high gain differential input device that increases
the magnitude of the applied signal to produce an
error voltage.
NOMINAL VALUE
The stated or objective value of a quantity or component,
which may not be the actual value measured.
NOMINAL VOLTAGE
The stated or objective value of a given voltage, which
may not be the actual value measured.
OPERATIONAL POWER SUPPLY

A power supply with a high open loop gain regulator which
acts like an operational amplifier and can be programmed
with passive components.
OPTO-COUPLER
A package that contains a light emitter and a photoreceptor used to transmit signals between electrically isolated
circuits.
OFF LINE POWER SUPPLY

1) A power supply in which the ac line is rectified and filtered without using a line frequency isolation transformer.
2) A power supply switched into service upon line loss to
provide power to the load without significant interruption.
See also UNINTERRUPTIBLE POWER SUPPLY.
OFFSET CURRENT
The direct current that appears as an error at either
terminal of a dc amplifier when the input current source
is disconnected.
SEC.4
OPTO-ISOLATOR
See OPTO-COUPLER.
78
OSCILLATOR
A nonrotating device for producing alternating current, the
output frequency of which is determined by the characteristics of the device. (IEEE Std 100-1988)
T E C H N I C A L
SEC.4
R E S O U R C E S
GLOSSARY
OUTPUT
The energy or information delivered from or
through a circuit or device.
page 79
OVERVOLTAGE
1) The potential difference between the equilibrium of
an electrode and that of the electrode under an imposed
polarization current. 2) A voltage that exceeds specified
limits.
OUTPUT CURRENT LIMITING

A protective feature that keeps the output current of a
power supply within predetermined limits during overload
to prevent damage to the supply or the load.
OVERVOLTAGE PROTECTION (OVP)

A feature that senses and responds to a high voltage
condition. See also OVERVOLTAGE, CROWBAR.
OUTPUT FILTER
One or more discrete components used to attenuate
output ripple and noise.
OVP
Abbreviation for Overvoltage Protection.
OUTPUT IMPEDANCE
The impedance that a power supply appears to present to
its output terminals.
PAD
A conductive area on a printed circuit board used for
connection to a component lead or terminal area, or
as a test point.
OUTPUT IMPEDANCE
The specified range over which the value of a stabilized
output quantity (voltage or current) can be adjusted.
OUTPUT RIPPLE AND NOISE
See PERIODIC and RANDOM DEVIATION.
PARALLEL
1) Term used to describe the interconnection of power
sources in which like terminals are connected such that
the combined currents are delivered to a single load.
2) The connection of components or circuits in a shunt
configuration.
OUTPUT VOLTAGE
The voltage measured at the output terminals
of a power supply.
OUTPUT VOLTAGE ACCURACY

The tolerance in percent of the output voltage
PARALLEL
The connection of two or more power sources of the same
output voltage to obtain a higher output current. Special
design considerations may be required for parallel operation of power sources.
OVERCURRENT DEVICE
A device capable of automatically opening an electric
circuit, both under predetermined overload and
short-circuit conditions, either by fusing of metal or by
electromechanical means.
PARD (periodic and random deviation):

Replaces the former term ripple of noise. PARD is the
periodic and random deviation referring to the sum of all
the ripple and noise components on the dc output of a
power supply regardless of nature or source
OVERCURRENT PROTECTION
See OUTPUT CURRENT LIMITING.
OVERLOAD PROTECTION
A feature that senses and responds to current of
power overload conditions. See also OUTPUT
CURRENT LIMITING.
PASS ELEMENT
A controlled variable resistance device, either a vacuum
tube or semiconductor, in series with the dc power source
used to provide regulation.
OVERSHOOT
A transient change in output voltage in excess of specified
output regulation limits, which can occur when a power
supply is turned on or off, or when there is a step change
in line or load.
PEAK
Maximum value of a waveform reached during a particular
cycle or operating time.
79
PEAK INVERSE VOLTAGE (PIV)

Maximum value of voltage applied in a reverse direction.
T E C H N I C A L
R E S O U R C E S
GLOSSARY
PEAK OUTPUT CURRENT

The maximum current value delivered to a load under
specified pulsed conditions.
SEC.4
page 80
POWER SOURCE
Any device that furnishes electrical power, including a
generator, cell, battery, power pack, power supply, solar
cell, etc.
PEAK-TO-PEAK
The measured value of a waveform from peak in a positive
direction to peak in a negative direction.
PERIODIC AND RANDOM DEVIATION (PARD)
The sum of all ripple and noise components measured
over a specified band width and stated, unless otherwise
specified, in peak-to-peak values.
POWER SUPPLY
A device for the conversion of available power of one set
of characteristics to another set of characteristics to meet
specified requirements.Typical application of power supplies include to convert raw input power to a controlled or
stabilized voltage and/or current for the operation of electronic equipment.
PIV
Abbreviation for Peak Inverse Voltage.
PPM
Abbreviation for parts per million.
POWER SUPPLY CORD

An assembly of a suitable length of flexible cord provided
with an attachment plug at one end.
PHASE ANGLE
The angle that a voltage waveform leads or lags the current waveform.
PREREGULATION
The initial regulation circuit in a system containing at least
two cascade regulation loops.
POLARITY
Property of device or circuit to have poles such as north
and south or positive and negative.
PRIMARY-SIDE-CONTROL
A name for an off-line switching power supply with the
pulse-width modulator in the primary.
POSITIVE RAIL
The most positive of the two output conductors
of a power supply.
PREREGULATION
A circuit electrically connected to the input or source of
power to the device.
POST REGULATION
Refers to the use of a secondary regulator on a power
supply output to improve line/load regulation and to
attenuate ripple and noise.
PROGRAMMABLE COEFFICIENT
The required range in control resistance to produce a
one volt change in output voltage. Expressed in ohms
per volt. The ratio of change in a control parameter to
induce a unit change in an output, e.g., 100 ohms/volt,
or 100 ohms/ampere.
POT
Abbreviation for potentiometer.
POTTING
An insulating material for encapsulating one or more
circuit elements
PROGRAMMABLE POWER SUPPLY

A power supply with an output controlled by an applied
voltage, current, resistance or digital code.
POWER FACTOR
The ratio of true to apparent power expressed as a
decimal, frequently specified as lead or lag of the current
relative to voltage.
PROGRAMMING
The control of a power supply parameter, such as output
voltage, by means of a control element or signal.
POWER FACTOR CORRECTION

1) Technique of forcing current draw to approach being
in-phase with the voltage in an ac circuit. 2) Addition of
capacitors to an inductive circuit to offset reactance.
POWER RATING
Power available at the output terminals of a power
source based on the manufacturers specifications.
PULSE-WIDTH MODULATION (PWM)

A method of regulating the output voltage of a switching
power supply by varying the duration, but not the frequency, of a train of pulses that drives a power switch.
80
T E C H N I C A L
R E S O U R C E S
GLOSSARY
SEC.4
page 81
PULSE-WIDTH MODULATOR (PWM)

An integrated discrete circuit used in switching-type power
supplies, to control the conduction time of pulses produced by the clock.
REGULATION
The process of holding constant selected parameters,
the extent of which is expressed as a percent.
PUSH-PULL CONVERTER
A power switching circuit that uses two or more power
switches driven alternately on and off.
REMOTE CONTROL
1) (general) Control of an operation from a distance: this
involves a link, usually electrical, between the control
device and the apparatus to be operated. Note: Remote
control may be over (A) direct wire, (B) other types of
interconnecting channels such as carrier-current or
microwave, (C) supervisory control, or (D) mechanical
means. 2) (programmable instrumentation) A method
whereby a device is programmable via its electrical
interface connection in order to enable the device to
perform different tasks. (IEEE Std 100-1988)
REGULATOR
The power supply circuit that controls or stabilizes the
output parameter at a specified value.
PUSH-PULL CIRCUIT
A circuit containing two like elements that operate in 180degree phase relationship to produce additive output components of the desired wave, with cancellation of certain
unwanted products. Note: Push-pull amplifiers and pushpull oscillators are examples. (IEEE Std 100-1988)
PWM
Variously, the abbreviation for Pulse-Width Modulation,
Pulse-Width Modulator
REMOTE PROGRAMMING
See PROGRAMMING.
REMOTE SENSING
A technique for regulating the output voltage of a power
supply at the load by connecting the regulator error-sensing leads directly to the load. Remote sensing compensates for specified maximum voltage drops in the load
leads. Care should be exercised to avoid opening load
handling leads to avoid damaging the power supply. Polarity must be observed when connecting sense leads to
avoid damaging the system.
RATED OUTPUT CURRENT

The maximum continuous load current a power supply is
designed to provide under specified operating conditions.
RECOVERY TIME
The time required for the measured characteristic to return
to within specified limits following an abnormal event.
RECTIFICATION
The process of changing an alternating current to a
unidirectional current. See FULL-WAVE RECTIFIER,
HALF-WAVE RECTIFIER.
REPEATABILITY
The ability to duplicate results under identical
operating conditions.
RECTIFIER
A component that passes current only in one direction,
e.g., a diode.
RESET SIGNAL
A signal used to return a circuit to a desired state.
RESISTANCE (R)
Property of a material that opposes the flow of current.
REFERENCE GROUND
Defined point in a circuit or system from which potential
measurements shall be made.
RESOLUTION
The smallest increment of change in output that can be
obtained by an adjustment.
REFERENCE VOLTAGE
The defined or specified voltage to which other voltages
are compared.
REGULATED POWER SUPPLY

A device that maintains within specified limits a constant
output voltage or current for specified changes in line, load
temperature or time.
81
RESONANCE
1) The state in which the natural response frequency of a
circuit coincides with the frequency of an applied signal, or
vice versa, yielding intensified response. 2) The state in
which the natural vibration frequency of a body coincides
with an applied vibration force, or vice versa, yielding reinforced vibration of the body.
T E C H N I C A L
SEC.4
R E S O U R C E S
GLOSSARY
RESONANT CIRCUIT
A circuit in which inductive and capacitive elements are in
resonance at an operating frequency.
page 82
RMS VALUE
In text, use lower case: rms. Abbreviation for Root Mean
Square Value.
RESONANT CONVERTER
A class of converters that uses a resonant circuit as part of
the regulation loop.
ROOT MEAN SQUARE (RMS) VALUE

1) (periodic function) The square root of the average
of the square of the value of the function taken
throughout one period (IEEE Std 100-1988).
2) For a sine wave, 0.707 x Peak Value.
RESONANT FREQUENCY
The natural frequency at which a circuit oscillates or a device vibrates. In an L-C circuit, inductive and capacitive reactances are equal at the resonant frequency.
RESPONSE TIME
The time required for the output of a power supply or
circuit to reach a specified fraction of its new value after
step change or disturbance.
SAFE OPERATING AREA (SOA)

A manufacturer specified power/time relationship
that must be observed to prevent damage to power
bipolar semiconductors.
RETURN
The name for the common terminal of the output of a
power supply; it carries the return current for the outputs.
SAFETY COMPLIANCE
Certification, recognition or approval by safety agencies
such as Underwriters Laboratories Inc. (UL/U.S.A.),
Canadian Standards Association (CSA), etc. See also
COMPLIANCE.
REVERSE VOLTAGE PROTECTION

A circuit or circuit element that protects a power supply
from damage caused by a voltage of reverse polarity
applied at the input or output terminals.
SAFETY GROUND
A conductive path from a chassis, panel or case to
earth to help prevent injury or damage to personnel
and equipment.
RFI
Abbreviation for Radio Frequency Interference.
SCR
Abbreviation for Silicon-Controlled Rectifier.
RIDE-THROUGH
See HOLDUP TIME
SECONDARY CIRCUIT
A circuit electrically isolated from the input or source of
power to the device.
RIPPLE
The periodic ac component at the power source output
harmonically related to source or switching frequencies.
Unless specified otherwise, it is expressed in peak-to-peak
units over a specified band width.
SECONDARY OUTPUT
An output of a switching power supply that is not
sensed by the control loop.
RIPPLE AND NOISE

See PERIODIC and RANDOM DEVIATION (PARD).See
PERIODIC and RANDOM DEVIATION (PARD).
SENSE AMPLIFIER
An amplifier which is connected to the output voltage
divider to determine, or sense, the output voltage.
RIPPLE VOLTAGE
The periodic ac component of the dc output of a
power supply.
SENSE LINE
The conductor which routes output voltage to the control
loop. See also REMOTE SENSING.
RISE TIME
The time required for a pulse to rise from 10 percent to 90
percent of its maximum amplitude.
82
SENSE LINE RETURN

The conductor which routes the voltage on the output
return to the control loop. See also REMOTE SENSING.
T E C H N I C A L
R E S O U R C E S
GLOSSARY
SEQUENCING
The process that forces the order of turn on and turn off of
individual outputs of a multiple output power supply.
page 83
SHORT CIRCUIT TEST

A test in which the output is shorted to ensure that the
short circuit current is within its specified limits.
SERIES
1) The interconnection of two or more power sources such
that alternate polarity terminals are connected so their
voltages sum at a load. 2) The connection of circuit components end to end to form a single current path.
SHUNT
1) A parallel conducting path in a circuit. 2) A low value
precision resistor used to monitor current.
SHUNT REGULATOR
A linear regulator in which the control element is in parallel
with the load, and in series with an impedance, to achieve
constant voltage across the load.
SERIES PASS
A controlled active element in series with a load that is
used to regulate voltage.
SI
Abbreviation for System International d'Unites.
SERIES REGULATOR
A regulator in which the active control element is
in series with the dc source and the load.
SIGNAL GROUND
The common return or reference point for analog signals.
SERIES REGULATION
See LINEAR REGULATION
SINE WAVE
A wave form of a single frequency alternating current
whose displacement is the sine of an angle proportional to
time or distance.
SETTING RANGE
The range over which the value of the stabilized output
quantity may be adjusted.
SLAVE
A power supply which uses the reference in another power
supply, the master, as its reference
SETTING TIME
The time for a power supply to stabilize within specifications after an excursion outside the input/output design
parameters.
SLEW RATE
The maximum rate of change a power supply output can
produce when subjected to a large step response or specified step change. The power supply is turned on.
SHIELD
Partition or enclosure around components in a circuit to
minimize the effects of stray magnetic and radio frequency
fields. See also ENCLOSURE, ELECTROSTATIC
SHIELD, FARADAY SHIELD.
SLOW START
A feature that ensures the smooth, controlled rise of the
output voltage, and protects the switching transistors from
transients when the power supply is turned on.
SHOCK HAZARD
A potentially dangerous electrical condition that
may be further defined by various industry or agency
specifications.
SNUBBER
An RC network used to reduce the rate of rise of voltage in
switching applications
SHORT CIRCUIT
A direct connection that provides a virtually zero
resistance path for current.
SOA
Abbreviation for Safe Operating Area.
SHORT CIRCUIT
The initial value of the current obtained from a power
source in a circuit of negligible resistance
SHORT CIRCUIT PROTECTION

A protective feature that limits the output current of a
power supply to prevent damage.
SEC.4
SOFT STARTS
Controlled turn on to reduce inrush currents.
83
SOURCE
Origin of the input power, e.g., generator, utility lines,
mains, batteries, etc.
T E C H N I C A L
SEC.4
R E S O U R C E S
GLOSSARY
SOURCE VOLTAGE EFFECT

The change in stabilized output produced by a specified
primary source voltage change.
TRANSIENT RESPONSE TIME

The room temperature or temperature of the still air surrounding the power supply, with the supply operating.
STABILITY
1) The percent change in output parameter as a function
of time, with all other factors constant, following a specified warm-up period. 2) The ability to stay on a given frequency or in a given state without undesired variation.
TEMPERATURE COEFFICIENT
The average percent change in output voltage per
degree centigrade change in ambient temperature
over a specified temperature range. See also AMBIENT
TEMPERATURE.
STANDOFF
A mechanical support, which may be an insulator, used to
connect and support a wire or device away from the
mounting surface.
TEMPERATURE DERATING
The amount by which power source or component
ratings are decreased to permit operation at elevated
temperatures.
STEP-DOWN TRANSFORMER
(power and distribution transformer) A transformer in which
the power transfer is from a higher voltage source circuit
to a lower voltage circuit. (IEEE Std 100-1988)
TEMPERATURE EFFECT
See TEMPERATURE COEFFICIENT.
STEP-UP TRANSFORMER
(power and distribution transformer) A transformer in which
the power transfer is from a lower voltage source circuit to
a higher voltage circuit. (IEEE Std 100-1988)
TEMPERATURE RANGE, OPERATING

See OPERATING TEMPERATURE RANGE
THERMAL PROTECTION
A protective feature that shuts down a power supply if its
internal temperature exceeds a predetermined limit.
STORAGE TEMPERATURE
The range of ambient temperatures through which an
inoperative power supply can remain in storage without
degrading its subsequent operation. See also AMBIENT
TEMPERATURE, OPERATING TEMPERATURE.
THREE TERMINAL REGULATOR

A power integrated circuit in a 3-terminal standard transistor package. It can be either a series or shunt regulator IC.
SUMMING POINT
The point at which two or more inputs of an operational
amplifier are algebraically added.
TIME CONSTANT
Time period required for the voltage of a capacitor in an
RC circuit to increase to 63.2 percent of maximum value
or decrease to 36.7 percent of maximum value.
SWITCHING FREQUENCY
The rate at which the dc voltage is switched in a converter
or power supply.
TOLERANCE
Measured or specified percentage variation from nominal.
SWITCHING FREQUENCY
A switching circuit that operates in a closed loop system to
regulate the power supply output.
TOTAL EFFECT
The change in a stabilized output produced by concurrent
worst case changes in all influence quantities within their
rated range.
SYNCHRONOUS RECTIFICATION
A rectification scheme in a switching power supply in
which a FET or bipolar transistor is substituted for the rectifier diode to improve efficiency.
SYSTEME INTERNATIONAL d'UNITES (SI)

The International System of Units comprised of Base
Units, Supplementary Units and Derived Units.
page 84
TRACE
A conducting path on a printed circuit board.
84
TRACKING
A characteristic of a multiple-output power supply that
describes the changes in the voltage of one output with
respect to changes in the voltage or load of another.
T E C H N I C A L
SEC.4
R E S O U R C E S
GLOSSARY
TRACKING REGULATOR
A plus or minus two-output supply in which one output
tracks the other.
page 85
UNINTERRUPTIBLE POWER SUPPLY (UPS

A type of power supply designed to support the load for
specified periods when the line varies outside specified
limits. See also OFF LINE POWER SUPPLY, ON LINE
POWER SUPPLY.
TRANSIENT
An excursion in a given parameter, typically associated
with input voltage or output loading.
UPS
Abbreviation for Uninterruptible Power Supply.
TRANSIENT EFFECT
The result of a step change in an influence quantity
on the steady state values of a circuit.
TRANSIENT RECOVERY TIME

The time required for the output voltage of a power supply
to settle within specified output accuracy limits following a
transient.
VARISTOR
A two electrode semiconductor device having a voltagedependent nonlinear resistance.
VDE
Abbreviation for Verband Deutscher Elektrotechniker.
TRANSIENT RESPONSE
Response of a circuit to a sudden change in an input or
output quantity.
VOLTAGE DIVIDER
Tapped or series resistance or impedance across a source
voltage to produce multiple voltages.
TRANSIENT RESPONSE TIME

The interval between the time a transient is introduced
and the time it returns and remains within a specified
amplitude range.
VOLTAGE DOUBLER
See VOLTAGE MULTIPLIER.
TTL
Abbreviation for transistor-transistor logic
VOLTAGE DROP
Difference in potential between two points in a passive
component or circuit.
VOLTAGE LIMIT
Maximum or minimum value in a voltage range.
UL
Abbreviation for Underwriters Laboratories Incorporated.
VOLTAGE LIMITING
Bounding circuit used to set specified maximum or minimum voltage levels.
UNDERSHOOT
A transient change in output voltage in excess of specified
output regulation limits. See OVERSHOOT.
VOLTAGE MODE
The functioning of a power supply so as to produce a stabilized output voltage.
UNDERVOLTAGE PROTECTION
A circuit that inhibits the power supply when output voltage
falls below a specified minimum.
VOLTAGE MONITOR
A circuit or device that determines whether or not an output voltage is within some specified limits.
UNDERWRITERS LABORATORIES INCORPORATED

(UL)
American association chartered to test and evaluate products, including power sources. The group has four locations so an applicant can interact with the office closest in
the country to his/her own location.
VOLTAGE MULTIPLIER
Rectifier circuits that produce an output voltage at a given
multiple greater than input voltage, usually doubling,
tripling, or quadrupling.
85
T E C H N I C A L
SEC.4
R E S O U R C E S
GLOSSARY
page 86
VOLTAGE REGULATION
The process of holding voltage constant between selected
parameters, the extent of which is expressed as a percent.
See also REGULATION.
WORST CASE CONDITION

A set of conditions where the combined influences on a
system or device are most detrimental.
VOLTAGE STABILIZATION
The use of a circuit or device to hold constant an output
voltage within given limits
X-RAY TUBE
A vacuum tube designed for producing X-rays by accelerating electrons to a high velocity by means of an electrostatic field and then suddenly stopping them by collision
with a target. (IEEE Std 100-1988)
VOLTAGE SOURCE
A power source that tends to deliver constant voltage.
VOLT (V)
Unit of measurement of electromotive force or potential
difference. Symbol E, in electricity; symbol V in semiconductor circuits.
ZENER DIODE
1) A diode that makes use of the breakdown properties of
a PN junction. If a reverse voltage across the diode is progressively increased, a point will be reached when the current will greatly increase beyond its normal cut-off value to
maintain a relatively constant voltage. Either voltage point
is called the Zener voltage. 2) The breakdown may be either the lower voltage Zener effect or the higher voltage
avalanche effect.
WARMUP
Process of approaching thermal equilibrium after turn on.
WARMUP DRIFT
The change in output voltage of a power source from turn
on until it reaches thermal equilibrium at specified operating conditions.
ZENER VOLTAGE
The reverse voltage at which breakdown occurs in a
zener diode.
WARMUP EFFECT
Magnitude of change of stabilized output quantities
during warmup time.
WARMUP TIME
The time required after a power supply is initially
turned on before it operates according to specified
performance limits.
WATT (W)
Unit of measure of power equal to 1 joule/sec. (W=EI)
WEBER (Wb)
The SI unit of magnetic flux equal to 108 maxwells. The
amount of flux that will induce 1 volt/turn of wire as the flux
is reduced at a constant rate to zero over a period of one
second.
WITHSTAND VOLTAGE
The specified operating voltage, or range of voltages, of a
component, device or cell.
WORKING VOLTAGE
The specified operating voltage, or range of voltages, of a
component, device or cell.
86
Spellman USA and Corporate HQ

475 Wireless Blvd.
Hauppauge, NY 11788
United States
tel: +1-631-630-3000
fax: +1-631-435-1620
Spellman Valhalla NY USA
One Commerce Park
Valhalla, NY 10595
United States
tel: +1-914-686-3600
fax: +1-914-686-5424
Spellman Bohemia NY USA
30 Crossways East
Bohemia, NY 11716
United States
Spellman UK
Broomers Hill Park #14, Broomers Hill
Pulborough, West Sussex,
United Kingdom RH20 2RY
tel: +44 (0) 1798 877000
fax: +44 (0) 1798 872479
Spellman Japan
4-3-1 Kamitoda,
Toda-shi, Saitama-ken,
Japan 335-0022
tel: +81(0) 48-447-6500
fax: +81(0) 48-447-6501
Spellman China
Spellman High Voltage Electronics (SIP) Co Ltd.
No. 5 Xinghan Street, Block A #2-03/08
Suzhou Industrial Park, China 215021
tel: +(86)-512-69006010
fax: +(86)-512-67630030
Spellman de Mexico SA de CV
Diagonal Lorenzo de la Garza # 65
Cd. Industrial
H. Matamoros, Tamps CP 87494
Mexico
tel: +52 868 150-1200
fax: +52 868 150-1218
Spellman de Mexico
Avenida Pedregal No. 2
Entre Avenida Chapultepec y Esquina
Parque Industrial Finsa
H. Matamoros, Tamps CP 87340
Mexico
www.spellmanhv.com
e-mail:sales@spellmanhv.com

High Voltage Reference Manual

Diunggah oleh

Informasi Dokumen

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

High Voltage Reference Manual

Diunggah oleh

Hak Cipta:

Format Tersedia

HIGH VOLTAGE REFERENCE MANUAL

SUPPLY CAN BE A BAD THING

IMPORTANT IN PROGRAMMING POWER SUPPLIES

SPECIFYING HIGH VOLTAGE POWER SUPPLIES

FREQUENTLY ASKED QUESTIONS

The only limit to the magnitude of short circuit current is

Are your supplies current protected?

Virtually all of Spellman's supplies (with the exception of a

When a short circuit is placed upon the output of a

Why is the short circuit repetition rate of my

In summary, the instantaneous short circuit current is a

How frequently a power supply is short circuited is an

What kind of high voltage connector do you

While most Spellman supplies typically come with one of

Many of our lower power modular supplies are provided

Supplies can be modified to enhance their short circuit

What is the difference between instantaneous short

The output section of a typical high voltage power supply

Most higher power units, both modular and rack mounted,

FREQUENTLY ASKED QUESTIONS

controlled and programmed via a PC software interface

is stripped back exposing the polyethylene insulation. The

What is a safe level of high voltage?

Below is a photo of a typical detachable high voltage

Typical Detachable High Voltage Cable

Can I program your supplies with a computer?

Where can I obtain information on high voltage

Yes, Spellman supplies can be programmed and

Most of Spellmans newer product releases come

One of the most comprehensive publications regarding

Many of our standard products that do not show the SIC

What is an "external interlock"? Why should I use it?

Virtually all of our products can be remote programmed

This simple circuit allows the customer to connect their

FREQUENTLY ASKED QUESTIONS

users, where a single specific usage needs to be

switch could be an interlock connection on a HV access

What is the difference between voltage

Voltage mode and current mode are the two regulating

What is the difference between a modular

Modular supplies and rack supplies are the two generic

Current mode works in a similar fashion, except it limits

Typically, rack mounted supplies are higher in power than

These two regulating modes work together to provide

Modular supplies tend to be lower power units (tens to

When ease of use and flexibility is required, like in a

What is power control? When would it be used?

Power control, (a.k.a. power mode or power loop) is a

FREQUENTLY ASKED QUESTIONS

power feedback signal (voltage x current = power). Using

Power control is typically used in two types of applications.

What is solid encapsulation?

The more popular usage of a power mode is in the area

What is floating ground?

Additionally, solid encapsulation provides the feature of

Whatever current flows out of the high voltage output of a

Why is oil insulation used?

Spellman has invested in and developed the use of oil

Essentially, the analog grounds inside the supply are

FREQUENTLY ASKED QUESTIONS

Corona is a luminous, audible discharge that occurs when

What is a voltage multiplier?

A voltage multiplier circuit is an arrangement of capacitors

Spellman has pioneered the use of voltage multiplier